Langrangian Hamilton

CHAPTER - I
LAGRANGE'S FORMULATION
Unit 1:
In mechanics we study particle in motion under the action of a force.
Equation of motion describes how particle moves under the action of a force.
However, every motion of a particle is not free motion, but rather it is restricted by
putting some conditions on the motion of a particle or system of particles. Hence the
basic concepts like equations of motion, constraints and type of constraints on the
motion of a particle, generalized coordinates, conservative force, conservation
theorems, DAlemberts principle, etc. on which the edifice of mechanics is built are
illustrated in this unit.
Introduction :
Mechanics is a branch of applied mathematics deals with the motion
of a particle or a system of particle with the forces Suppose a bullet is fired from a
fixed point with initial velocity u, not exactly vertically upward but making an angle
y
with the horizontal. Then
what instruments do
mathematicians need to find the
. position of the bullet at some
y = u sin u .
y = u sin - gt instant latter, its velocity at that
instant, the distance covered by
. the bullet at that instant and also

x = u cos
the path followed by the bullet at
mg
x
O . the end of its journey?
x = u cos
Classical Mechanics Page No. 1

Well, to answer such questions, mathematicians do not need any meter stick
to measure the distance covered by the bullet at the instant, they don't need any
speedometer to find its speed at any instant t, nor they need any clock to see the time
required to cover the definite distance. In fact, they need not have to do any such
experiment. What they need to describe the motion of the bullet are simply the co-
ordinates. Hence the single most important notion in mechanics is the concept of co-
ordinates. But the co-ordinates however, just play a role of markers or codes and will
no way influence or affect the motion of the bullet. These are just mathematical tools
in the hands of a mathematician. Thus the instruments in the hands of a
mathematician are the co-ordinates. With the help of these co-ordinates, the motion
of a particle or system of particle can completely be described.
For instance, to discuss the motion of the bullet, take P(x, y) be any point on
the path of the bullet. The only force acting on the bullet is the gravitational force in
the downward direction. Resolving this force horizontally and vertically, we write
from Newton's second law of motion the equations of motion as
x = 0,

. . . (1)
y = g.

Integrating the above two equations and using the initial conditions we
readily obtain
x = u cos ,
. . . (2)
y = u sin gt ,
where u is the initial velocity of the bullet when t = 0. Integrating equations (2) once
again and using the initial conditions we obtain
x = u cos t , . . . (3)
1 2
y = u sin t gt . . . . (4)
2

Equations (2) determine the velocity of the bullet at any time t, while equations (3)
and (4) determine the position of the bullet at that instant. Further, eliminating t from
equations (3) and (4), we get
1 gx 2
y = x tan . . . . (5)
2 u 2 cos 2
This equation gives the path of the bullet and the path is a parabola.
However, the co-ordinates used to describe the motion of a particle or system
of particles must be linearly independent. If not then the number of equations
describing the motion of the system will be less than the number of variables and in
this case the solution can not be uniquely determined. For example if the particle
moves freely in space, then three independent co-ordinates are used to describe its
motion. These are either the Cartesian co-ordinates (x, y, z) or the spherical polar co-
ordinates ( r , , ) . If however, the particle is moving along one of the co-ordinate
axes in space, then all the three co-ordinates are not independent, hence these three
co-ordinates can not be used for its description. Along a co-ordinate axis only one
co-ordinate varies and other two are constants and only the varying co-ordinate can
be used to describe the motion of the particle.
Basic concepts:
1. Velocity: Let a particle be moving along any path with respect to the fixed
point O. If r is its position vector, then the velocity of the particle is defined as the
time rate of change of position vector. i.e.,
v = r ,
where dot denotes the derivative with respect to time. If further r = xi + yj + zk is
the position vector, then velocity of the particle is v = r = xi

+ yj
+ zk
, where x , y , z
are called the components of the velocity along the coordinates axes.

2. Linear momentum: The linear momentum of a particle is defined as the
product of mass of the particle and its velocity. It is a vector quantity and is denoted
by p. Thus we have p = mv . The direction of momentum is along the same direction
of the velocity. In terms of the linear momentum of the particle the equation of
motion is given by F = p .
3. Angular momentum: The angular momentum of a particle about any fixed
point O as origin is defined as r p . It is a vector quantity and denoted by L . Thus
L = r p . Angular momentum is perpendicular to both the position vector and the

linear momentum of the particle.
4. Torque (Moment of a Force): The time rate of change of angular
momentum L is defined as torque, It is denoted by N . Thus
dL d
N= = ( r p)
dt dt
d
N = ( r mr ) ,
dt
= r mr + r F,
dL
N= = r F.
dt
Equation of Motion and Conservation Theorems :
1. For a Particle :
Consider a particle of mass m whose position vector with respect to some
fixed point is r . If F is a force applied on the particle then the equation of motion
of the particle is given by Newtons second law of motion
dp
F= , . . . (1)
dt
where
dr
p=m = mr
dt

is the linear momentum of the particle. The force is defined to be
F = mass accel.
F = ma .
Hence equation (1) becomes
d 2r
= a. . . . (2)
dt 2
Integrating this equation we get
dr
= at + c , . . . (3)
dt
where c is the constant of integration and is to be determined. Now applying the
dr
initial conditions, we have when t = 0, = u initial velocity.
dt
c=u
v = u + at . . . . (4)
This equation determines the velocity of the particle at any instant t. Integrating (4)
again we get
1
r = ut + at 2 + c1 ,
2
where c1 is the constant of integration. At t = 0, r = 0 c1 = 0 . Hence we have
1
r = ut + at 2 . . . . (5)
2
This equation gives the distance covered by the particle at any time t. One can
combine equations (4) and (5) and write
v 2 = u 2 + 2ar . . . . (6)
This equation determines the velocity of the particle at a given distance. Equations
(4), (5) and (6) are the algebraic equations of motion and are derived from the
equation (1) namely

F = p . . . . (7)
This is the differential equation of motion. It follows from equation (7) that if the
applied force is zero then the linear momentum of the particle is conserved.
Equation of motion of a system of particles:

Consider a system of n particles of masses m1 , m2 ,..., mn having position
vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. Any particle of this

system will experience two types of forces.
i) External forces on the system Fi ( e ) , i = 1, 2,..., n .
ii) Internal forces Fji( int ) .
Thus the total force acting on the i th particle of the system is given by
Fi = Fi ( e ) + Fji( int ) ,
j
where F( j
ji
int )
is the total internal force acting on the i th particle due to the
interaction of all other (n-1) particles of the system. Thus the equation of motion of
the i th particle is given by
Fi ( e) + Fji( int ) = p i . . . . (1)
j
The equation of motion of the whole system is obtained by summing over i the
equation (1) we get

F( ) + F(
i
e
ji
int )
= p i .
i i j i
We write this equation as
F( ) +
i
i
e
i , j ,i j
Fji( int ) = p i .
i
. . . (2)

The term
i , j ,i j
Fji( int ) represents the vector sum of all the interaction forces due to the
presence of remaining n-1 particles. However, there is no self interacting force,

hence Fji = 0, i = j . Also the internal forces obey the Newtons third law of
motion. That is the action of one particle on the other is equal but opposite to the
action of second on the first. This implies that the mutual interaction between the
i th and j th particles are equal and opposite. i.e.
Fji(int ) = Fij( int ) .
This gives
i , j ,i j
Fji( int ) = 0 .
Thus equation of motion (2) of the system becomes
F ( ) = p ,
i
i
e
i
i
F e = P , . . . (3)
where P is the total momentum of the system and F e = P is the total external force
acting on the system.
Conservation Theorem of Linear momentum of the system of particles :
Theorem 1 : If the sum of external forces acting on the particles is zero, the total
linear momentum of the system is conserved.
Proof : Proof follows immediately from equation (3). i.e., if
F e = 0 P = const.
Angular Momentum of the system of particles :
Consider a system of n particles of masses m1 , m2 ,..., mn having position
vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. The angular momentum of
the i th particle of the system about the origin is given by

Li = ri pi .

Thus the total angular momentum of the system about any point is equal to the vector
sum of the angular momentum of individual particles. Hence we have
L = r p .
i
i
i
i i . . . (1)
If N is the total torque acting on the system, then equation of motion of the system is
given by
dL d
N= = ri pi .
dt dt i
dL
N= = ri pi + ri p i . . . . (2)
dt i i
But we have
r p = r m r = 0 .
i
i i
i
i i i . . . (3)
Now consider

r p = r F ( ) + F (
i i i i
e
ji
int )

i i j
r p = r F ( ) + r F (
i
i i
i
i i
e
i
i
j
ji
int )
r p = r F ( ) + r F (
i
i i
i
i i
e
i, j
i ji
int )
. . . (4)
However, the term r F(

i, j
i ji
int )
can be expanded for i j as
r F(
i, j
i ji
int )
= ( r2 r1 ) F12( int ) + ( r3 r1 ) F13( int ) + ( r3 r2 ) F23( int ) + ...
r F(
i, j
i ji
int )
= ri rj Fji( int ) ,
r F(
i, j
i ji
int )
= rij Fji( int ) , for rij = ri rj . . . (5)
Interchanging i and j on the r. h. s. of equation (5) we get

r F(
i, j
i ji
int )
= rji Fij(int ) ,
r F(
i, j
i ji
int )
= rij Fji( int ) . . . (6)
Adding equations (5) and (6) we get
r F(
i, j
i ji
int )
=0 . . . (7)
Consequently, on using equations (3), (4) and (7) in equation (2) we readily obtain
dL
= ri Fi ( ) .
e
N= . . . (8)
dt i
This equation shows that the total torque on the system is equal to the vector sum of
torques acting on the individual particles of the system.
Conservation Theorem of Angular momentum of the system of particles:
Theorem 2 : If the total external torque acting on the system of particles is zero,
then the total angular momentum of the system is conserved.
Proof : Proof follows immediately from equation (8). i.e., if
N = 0 L = const.
Some definitions:
Centre of Gravity (Centre of Mass): It is the point of the body at which the whole
mass of the body is supposed to be concentrated. If R is the position vector of the
centre of mass of the body with respect to the origin then its coordinates are given by
m r i i
R = ( x, y ) = i
, where M = mi is the total mass of the body.
M i
Example 1: Show that the total angular momentum of a system of particles can be
expressed as the sum of the angular momentum of the motion of the centre of mass
about origin plus the total angular momentum of the system about the centre of mass.

Solution: Consider a system of n particle of masses m1 , m2 ,..., mn having position
vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. The angular momentum of
the i th particle of the system about the origin is given by

Li = ri pi .
Thus the total angular momentum of the system about any point is equal to the vector
sum of the angular momentum of individual particles. Hence we have
L = r p .
i
i
i
i i (1)
Let R be the radius vector of the centre of mass with respect to the origin and ri be
the position vector of the i th particle with respect to the centre of mass. Then we have
ri = ri + R. (2)
Differentiating this equation with respect to t we get
ri = ri + R .
i.e., vi = vi + v,
where vi is the velocity of the i th particle with respect to O,
vi - velocity of the i th particle with respect to centre of mass,
v - velocity of the centre of mass with respect to O.

Using the equation (2) in the equation (1) we get
(
L = Li = ri + R mi ( v + vi ) ,
i i
)
L = ri mi v + mi ri v + R mi v + R mi vi,
i i i i
d
L = ri pi + mi ri v + R mi v + R mi ri,
i i i dt i
Consider the term
m r = m ( r R ) ,
i
i i
i
i i

m r = m r m R,
i
i i
i
i i
i
i
m r = MR MR ,
i
i i
m r = 0.
i
i i
Consequently we have from above equation
L = ri pi + R mi v,
i i
L = R Mv + ri pi. (4)
i
This shows that the total angular momentum about the point O is the sum of the
angular momentum of the centre of mass about the origin and the angular momentum
of the system about the centre of mass.
Constraint Motion :
Some times the motion of a particle or a system of particles is not free but it
is limited by putting some restrictions on the position co-ordinates of the particle or
system of particles. The motion under such restrictions is called constraint motion or
restricted motion. The mathematical relations establishing the limitations on the
position co-ordinates are called as the equations of constraints. Mathematically, the
constraints are thus the relations between the co-ordinates and the time t.
Consequently, all the co-ordinates are not linearly independent; constraint relations
relate some of them. Thus in general the constraints on the motion of a particle or
system of particles are always possible to express in the form
f r ( xi , yi , zi , t ) or or = 0 ,
where r =1,2,3,...,k the number of constraints and ( xi , yi , zi ) are the position co-
ordinates of the i th particle of the system, i = 1, 2,..., n .

Examples of motion under constraints:
1. The motion of a rigid body,
2. The motion of a simple pendulum,
3. The motion of a particle on the surface of a sphere,
4. The motion of a particle along the parabola x 2 = 4ay ,
5. The motion of a particle on an inclined plane.
Holonomic and non-holonomic Constraints:
If the constraints on a particle or system of particles are expressible as
equations in the form
f r ( xi , yi , zi , t ) = 0, r = 1, 2,..., k . . . (1)
then constraints are said to be holonomic otherwise non-holonomic constraints. A

system of particles is called respectively holonomic or non-holonomic system if it
involves holonomic or non-holonomic constraints.
For example:
Constraints involved on the motion of a rigid body, and simple pendulum are
examples of holonomic constraints, while constraints involved in the motion of a
particle on the surface of a sphere, the motion of a gas molecules inside the container
are the examples of non-holonomic constraints. However, this is not the only way to
describe the non-holonomic system. A system is also said to be non-holonomic, if it
corresponds to non-integrable differential equations of constraints. Such constraints
can not be expressed in the form of equation of the type
f r ( xi , yi , zi , t ) = 0 .
Hence such constraints are also called non-holonomic constraints. Obviously,

holonomic system has integrable differential equations of constraints expressible in
the form of equations.
The constraints are further classified in to two parts viz., Scleronomic and
Rheonomic Constraints.

Scleronomic and Rheonomic Constraints :
When the constraint relations do not explicitly depend on time are called
scleronomic constraints. While the constraints, which involve time explicitly are
called rheonomic constraints. The examples cited above are all scleronomic
constraints. A bead moving along a circular wire of radius r with angular velocity
is an example of rheonomic constraint and the constraints relations are
x = r cos t , y = r sin t .
Worked Examples
Example 2 : Consider a system of two particles joined by a mass less rod of fixed
length l . Suppose for simplicity, the system is confined to the horizontal plane xy .
Suppose further that the system is so constrained that the centre of the rod cannot
have a velocity component perpendicular to the rod. Show that the constraint
involved in the system is non-holonomic.
y Solution: Let (x1, y1) and (x2, y2) be the
positions of the two particles connected by
v=0
the mass less rod of length l . The system is
(x2 , y2 ) shown in the fig.

Since the length between the two particles
x1 + x2 y1 + y2
,
(x1, y1 ) 2 2
is constant, clearly one of the constraint
x relations is
O
2 2
( x1 x2 ) + ( y1 y2 ) = l2 . . . (1)
x2 x1 = l cos ,
where, . . . (2)
y2 y1 = l sin
The constraint (1) is clearly holonomic. The other constraint is such that the
centre of the rod cannot have velocity component perpendicular to the rod.
Mathematically this is expressed as

( x1 + x2 ) cos ( 90 + ) + ( y1 + y 2 ) cos = 0 ,
( x1 + x2 ) sin = ( y1 + y2 ) cos . . . . (3)
This constraint can not be integrated and hence the constraint is non-holonomic and
consequently, the system is non-holonomic.
Degrees of freedom and Generalized co-ordinates :
Consider the motion of a free particle. To describe its motion we need three
independent co-ordinates, such as the Cartesian co-ordinates x, y, z or the spherical
polar co-ordinates r , , etc. The particle is free to execute motion along any one of
the axes independently with change in only one co-ordinate. In this case we say that
the particle has three degrees of freedom. Thus we define
Definition : The least possible number of independent co-ordinates required to

specify the motion of the system completely by taking into account the constraints is
called degrees of freedom.
e.g. For a system of N particles free from constraints moving independent of each
other has 3N degree of freedom.
Generalized co-ordinates:
A system of N particles free from constraints has 3N degrees of freedom. If
however, there exists k holonomic constraints expressed in k equations
f i ( r1 , r2 ,..., rn , t ) = 0, i = 1, 2,..., k , . . . (1)
then 3N co-ordinates are not all independent but related by k equations given in (1).
We may use these k equations to eliminate k of the 3N co-ordinates, and we are left
with 3 N k = n (say) independent co-ordinates. These are generally denoted by
qj, j = 1, 2,..., n called the generalized co-ordinates and the system has 3N-k
degrees of freedom.

Definition: A set of linearly independent variables q1 , q2 , q3 ,...qn that are used to
describe the configuration of the system completely by taking into account the
constraints forces acting on it is called generalized co-ordinates.
Thus in general we have
No. of degrees of freedom No. of constraints= No. of generalized co-ordinates.
Note: The generalized co-ordinates need not be the position co-ordinates, which have
the dimensions of length, breath and height, but they can be angles, charges or
momentum of the particle.
Transformation Relations:
It is always possible to express the position co-ordinates of a particle or a
system of particles in terms of generalized co-ordinates and vice-versa. This
expression is called the transformation relation.
e.g., If ri , i = 1, 2,3,...n are the position vectors of the n particles of the system and
q j , j = 1, 2,..., n are the generalized co-ordinates, then there exists a relation
ri = ri ( q1 , q2 , q3 ,...qn , t ) , . . . (1)
called the transformation relation.

Work: Let a force F be acted on a particle whose position vector is r . Suppose the
particle is displaced through an infinitesimal distance dr due to the application of
force F. Then the work done by the force F is given by
dW = F dr
If the particle is finitely displaced from point P ( r1 ) to P ( r2 ) along any path, then
the work done by F is given by

r2
W = F dr . . . (1)
r1
Conservative Force : The work given in expression (1) is in general depends on the
extreme positions of the particle and also the path along which it travels. If a force is

such that the work depends only upon the positions P1 , P2 and not on the path
followed by the particle, then the force F is called conservative force, otherwise non-
conservative.
Worked Examples
Example 3 : Show that the gravitational force is conservative.

Solution : Let a particle of mass m move along a curve PQ under gravity. Thus the
z only force acting on the particle is its own
weight in the down ward direction. Therefore,
Y
P
work done by the force is given by
Q
X
Q W = F dr ,
P
W = mg If F = Xi + Yj and r = xi + yj , where X and Y

are the components of the force along the co-
x
O b a ordinate axes. We see that X = 0 , and Y = w .
b
Hence W = wdy ,
a
where a and b are the ordinates at points P and Q respectively.

This implies
W = w( a b) .
This shows that the work does not depend upon the path but depends on the
extreme points. Hence the gravitational force is conservative. Alternately, we say
that, the force F is conservative if the work done by it around the closed path is zero.
i. e. F is conservative iff F dr = 0 . . . (1)
However, by Stokes theorem, we have
F dr = F .ds,
s
. . . (2)

where ds is an arbitrary surface element. Thus from equations (1) and (2) we have
F is conservative iff F = 0 . . . (3)
However, F = 0 F is a gradient of some potential V.
V
F = V or F = ,
r
where V is a potential called potential energy of the particle and is a function of
position only. Thus the force F is conservative if
F = V . . . (4)
and conversely. The negative sign indicates that F is in the direction of decreasing V.
Example 4: Show that the inverse square law of attractive force (central force) is
conservative.
Solution: The inverse square law of force is the force of attraction between two
particles and is given by
m1m2
F = G . . . (1)
r2
where negative sign indicates that the force is directed towards the fixed point and it
is called the attractive force. We write the force as
k
F = r, for k = Gm1m2 . . . (2)
r3
For r = xi + yj + zk
k ( xi + yj + zk )
We have F = 3
. . . (3)
(x 2 2
+y +z )
2 2
For conservative force we have F = 0.

Consider therefore
i j k

F =
x y z
Kx Ky Kz
r3 r3 r3
z y x z y x
F = K i 3 3 + j 3 3 + k 3 3 .
y r z r z r x r x r y r
3 yx 3 yx
3 yz 3 yz 3 xz 3 xz
F = K i 5 + 5 + j 5 + 5 + k 5 + 5 .
2 2 2
r r2 r r2 r r 2
F = 0.
This shows that the inverse square law of attractive force F is conservative.
Virtual Work :
If the system of forces acting on a particle be in equilibrium then their
resultant is zero and hence the work done is zero.
Thus in the case of a particle be in equilibrium there is no motion, hence there
arises no question of displacement. In this case we assume the particle receives a
small virtual displacement (the displacement of the system which causes no real
motion is called as virtual or imaginary displacement) and it is denoted by ri .
Virtual displacement ri is assumed to take place only in the co-ordinates and at

fixed instant t, hence change in time t is zero.
ri = ( dri )dt =o .
Thus the work done by the system of forces in causing imaginary
displacement is called virtual work. It is the amount of work that would have been

done if the actual displacement had been caused. Hence the expression for the virtual
work done by the forces is given by
Virtual work W = Fi ri . . . . (1)
i
Principle of Virtual Work :

If the forces are in equilibrium then the resultant is zero. Hence the algebraic
sum of the virtual work is zero. Conversely, if the algebraic sum of the virtual work
is zero then the forces are in equilibrium.
Note that this principle is applicable in statics. However, an analogous
principle in dynamics was put forward by DAlembert.
DAlemberts Principle :
DAlembert started with the equation of motion of a particle Fi = p i , where
pi is the linear momentum of the i th particle. This can be written as Fi p i = 0 .
Hence ( F p ) = 0,
i
i i
implying a system of particles is in equilibrium. This equation states that the

dynamical system appears to be in equilibrium under the action of applied forces Fi
and an equal and opposite effective forces p i . In this way dynamics reduces to
static. Thus
( F p ) = 0
i
i i the system is in equilibrium (the resultant is zero).
Hence the virtual work done by the forces is zero. This implies that
( F p ) r = 0 .
i
i i i
This is known as the mathematical form of DAlembert principle. This states that a
system of particles moves in such a way that the total virtual work done by the
applied forces and reverse effective forces is zero.

Note :
1. DAlembert Principle describes the motion of the system by considering its
equilibrium.
2. All the laws of mechanics may be derived from this single principle. Hence
DAlembert principle has been called the fundamental principle of
mechanics. We will solve some examples by using this principle.
Worked Examples
Example 5: A particle is constrained to move in a circle in a vertical plane xy. Apply

the DAlemberts principle to show that for equilibrium we must
xy
have yx gx = 0 .
Solution: Consider a particle of mass m be moving along
a circle of radius r in xy plane. Let (x, y) be the position
of the particle at any instant t with respect to the fixed- O
point 0.The constraint on the motion of the particle is that

(x, y)
the position co-ordinates of the particle always lie on the
circle. Hence the equation of the constraint is
W = mg
x2 + y 2 = r 2 . . . (1)
2 x x + 2 y y = 0
y
or x= y . . . (2)
x
where x and y are displacement in x and y respectively. Now from
DAlemberts principle, we have
( F mr) r = 0 .
In terms of components we have
( Fx mx) x + ( Fy my) y = 0 . . . . (3)

However, the only force acting on the particle at any instant t is its weight mg in the
downward direction. Resolving the force horizontally and vertically, we have Fx = 0
and Fy = mg . Therefore equation (3) becomes
mx x ( mg + my) y = 0 .
On using (2) we have

m ( yx + gx ) x = 0 .
xy +
For x 0, and m 0 we have

xy
yx gx = 0 , . . . (4)
which is the required equation of motion.
Example 6: Use DAlemberts principle to determine the equation of motion of a

simple pendulum.
Solution : Consider a particle of mass m attached to one end of the string and other
O x end is fastened to a fixed point 0. Let l be
the length of the pendulum and the
l
angular displacement of the pendulum
B(x, y) shown in the fig.
A According to the DAlemberts principle we
mg sin
mg mg cos have
y
( F p ) r = 0
i
i i i
where i is the number of particles in the system.

( F mr) r = 0 ,
where r is the distance of the particle from the starting point along the curve.
Resolving the force acting on the particle along the direction of motion and
perpendicular to the direction of motion we have
( mg sin mr) r = 0 ,
where the negative sign indicates the force is opposite to the direction of motion.
Since r 0 we have
r = g sin .
. . . (1)
From the figure we have r = arcAB r = l

r = l .
Equation (1) becomes
g
= sin . . . . (2)
l
g
For small angle, we have sin = = .
l
Generalized Velocities :
From transformation equations we have
ri = ri ( q1 , q2 , q3 ,...qn , t ) , . . . (1)
Differentiating this with respect to t we get

ri r
r = q j + i . . . (2)
j q j t
where q j , j = 1, 2,3,..., n are called generalized velocities.
Virtual displacement :
We find variation (change) in the transformation equation (1) to get
ri
ri = qj
q j
Note here that t term is absent because virtual displacement is assumed to take
place at fixed instant t, hence t = 0 .

Generalized force :
If Fi are forces acting on a dynamical system with position vectors ri then
virtual work done by these forces is given by
W = Fi ri ,
i
ri
= Fi qj ,
i j q j
r
= Fi i q j ,

j i q j
= Q j q j ,
j
where
ri
Q j = Fi . . . (1)
i q j
are called the components of generalized forces.

Note :
1. If forces are conservative then they are derived from potential V and are
given by
V
Fi = iV = .
ri
V
Consequently, the generalized forces are given by Q j = .
q j
2. If the forces are non-conservative, the scalar potential U may be function of

position, velocity and time. i.e., U = U ( q j , q j , t ) . This is called velocity dependent
potential or generalized potential. Such a potential exits in the case of a motion of a

particle of charge q moving in an electromagnetic field. We will see latter in example
(8) that how the generalized potential can be determined in the case of a particle
moving in an electromagnetic field. In this case generalized forces are given by

U d U
Qj = + .
q j dt q j
3. If however, the system is acted upon by conservative forces Fi and non-
conservative forces Fi ( d ) , in this case generalized forces Q j are given by
(
Q j = Fi + Fi (
d)
) qr
i
j
Qj =
V
q j
+ Q(j ) ,
d
where
ri
Q (j ) = Fi (
d d)
i q j
are non-conservative forces which are not derivable from the potential V. Such a
situation often arises when frictional forces or dissipative forces are present in the
system.
It is found by experiment that in general the dissipative or frictional forces are
proportional to the velocity of the particle.
Fi ( d ) = i ri ,
where i are constants. In such cases the generalized forces are obtained as
ri
Q(j d ) = Fi ( d ) ,
i q j
ri
= i ri
i q j
However, from transformation equation we obtain
ri r
= i .
q j q j
Thus we write
1 2
Q(j d ) = i ri ,
q j 2
R
= ,
q j
1
where R=
2
i ri 2
is called the Rayleighs dissipative function.
Unit 2: Lagranges Equations of motion:

Newtonian approach for the description of particle involves vector quantities.
We now introduce another formulation called the Lagrangian formulation for the
description of mechanics of a particle or a system of particles in terms of generalized
coordinates, generalized velocities with time t as a parameter. This formulation
involves scalar quantities such as kinetic energy and the potential energy and hence
proves to be easier then the Newtonian approach, because to deal with scalars is easy
than to deal with vectors.
Lagranges Equations of motion from DAlemberts Principle :
Theorem 3 : Obtain Lagranges equations of motion from DAlemberts principle.
Proof : Consider a system of n particles of masses mi and position vectors ri . We
know the position vectors ri are expressed as the functions of n generalized co-
ordinates q1 , q2 , q3 ,..., qn and time t as
ri = ri ( q1 , q2 , q3 ,...qn , t ) , . . . (1)
If Fi are the forces acting on the system, then by DAlemberts principle we have
( F p ) r = 0 ,
i
i i i . . . (2)
ri is the linear momentum of the i th particle of the system . From the

where, p i = mi
transformation equations we obtain the expression for the virtual displacement
ri
ri = qj ,
q j

where the term t is absent because the virtual displacement is assumed to take
place only in the co-ordinates and at the particular instant. Hence equation (2)
becomes
ri ri
F q q = m r q
i j
i j
i j
i i qj .
j j
ri r
F q i q j = mi
ri i q j ,
q j
j i j i, j
ri
or Q q = m r q
j
j j
i, j
i i qj , . . . (3)
j
where
ri
Q, = Fi . . . . (4)
q j
are called the components of generalized forces.

Consider
d ri r d r
ri =
ri i + ri i .
dt q j q j dt q j
Substituting this in equation (3) we get
d r d r
j j j
Q q = mi ri i ri i q j . . . . (5)
i, j dt q j dt q j
Now from equation (1) we have
ri r
ri = qk + i . . . . (6)
k qk t
Differentiating this with respect to q j we get
ri r
= i . . . . (7)
q j q j
Further, differentiating equation (6) w. r. t. q j , we get

ri 2 ri 2 ri
= qk + . . . . (8)
q j k qk q j t q j
Also we have
d ri 2 ri 2 ri
= qk + . . . . (9)
dt q j k q j qk q j t
We notice from equations (8) and (9) that
ri d r
= i
q j dt q j
In general we have
d d
= . . . . (10)
q j dt dt q j
On using equation (10) in equation (5) we get
d vi v
Q q = dt m v
j j i i
q j
mi vi i q j .
q j
j i, j

We write this as
d 1 1 2

Q q = dt q 2 m v
j j i i
2
q 2 mi vi q j ,

j j j i j i
d T T
or Q q = dt q
j j

qj ,
j j
j q
j
1
where T=
2 i
mi vi 2
is the total kinetic energy of the system of particles.

d T T
j dt q Qj q j = 0 . . . . (11)
j q j

If the constraints on the motion of particles in the system are holonomic then q j are
independent. In this case we infer from equation (11) that
d T T
Qj = 0 ,
dt q j q j
d T T
= Qj , j = 1, 2, 3,...., n . . . . (12)
dt q j q j
These are called the Lagranges equations of motion. We see that, to derive
the Lagranges equations of motion the knowledge of forces acting on the system of
particles will not be necessary.
Note : If the constraints are non-holonomic then the generalized co-ordinates are not
all independent of each other. Hence we cant conclude equation (12) from equation
(11).
Note: In deriving Lagranges equations of motion the requirement of holonomic
constraints does not appear until the last step.
Case (1) : Conservative system :
If the system is conservative so that particles move under the influence of a
potential which is dependent on co-ordinates only, then the forces are derived from
the potential V given by
V
Fi = iV = .
ri
In this case the components of generalized forces becomes
V ri V
Qj = = , and V V ( q j ) .
i ri q j q j
d (T V ) (T V )
= 0.
dt q j q j
Define a new function L = T V ,

where L which is a function of q1 , q2 , q3 ,..., qn , q1 , q2 , q3 ,..., qn and time t is called a
Lagrangian function of the system of particles. Then the equations of motion become
d L L
= 0 . j = 1, 2, ..., n . . . (13)
dt q j q j
These are called the Lagranges equations for motion for conservative holonomic
system.
Note : The Lagrangian L satisfying equation (13) is not unique. Refer Example (13)
bellow.
Case (2) : Non-conservative system :
In the case of non-conservative system the scalar potential U may be
function of both position and velocity. i.e., U = U ( q j , q j , t ) . Such a potential is
called as velocity dependent potential. In this case the associated generalized forces
are given by
U d U
Qj = + .
q j dt q j
Substituting this in the equation (12) we get
d L L
= 0 , j = 1, 2,..., n
dt q j q j
which are the Lagranges equations of motion for non-conservative forces.
Case (3) : Partially conservative and partially non-conservative system :
Consider the system is acted upon by conservative forces Fi and non-
conservative forces Fi ( d ) . Such a situation often occurs when frictional forces or

dissipative forces are present in the system. In this case the components of
generalized force are given by

(
Q j = Fi + Fi (
d)
) qr i
j
Qj =
V
q j
+ Q(j ) ,
d
where the non-conservative forces which are not derivable from potential function V
are represented in Q(j d ) . Substituting this in equation (12) we readily obtain
d L L
= Q (j ) , j = 1, 2,..., n
d
. . . (14)
dt q j q j
where the Lagrangian L contains the potential of the conservative forces, and Q(j d )
represents the forces not arising from the potential V. However, it is found by
experiment that, in general the dissipative or frictional forces are proportional to the
velocity of the particles.
Fi ( d ) = i ri , i are constants.
Hence we have
ri
Q(j d ) = Fi ( d ) ,
i q j
ri
= i ri .
i q j
But we know that

ri r
= i .
q j q j
Hence
1 2 R
Q (j d ) = i ri = ,
q j 2 q j
1
where R=
2 i
i ri 2
is called Rayleighs dissipation function. Hence the Lagranges equations of motion

become

d L L R
+ = 0. . . . (15)
dt q j q j q j
Worked Examples
Conservation of Energy:
Example 7: Show that the total energy of a particle moving in a conservative force
field remains constant, if the potential energy is not an explicit function of time.
Solution : Let a particle of mass m be moving in the conservative field of force F.
Let r be the position vector of the particle at any instant. The total energy of the
particle is
E = T +V , . . . (1)
where T = kinetic energy,
V = potential energy.
Differentiating (1) with respect to t we get
dE dT dV
= + , . . . (2)
dt dt dt
where the force
dv
F =m
dt
Therefore
dv dr
Fdr = m dt
dt dt
Fdr = mvdv,
1
Fdr = d mv 2
2
Fdr = dT ,
dr dT
F = . . . . (3)
dt dt

Similarly, we have the potential energy V =V(r, t), therefore,
V V V V
dV = dx + dy + dz + dt ,
x y z t
V
dV = V .dr + dt ,
t
dV dr V
= V . + . . . . (4)
dt dt t
dE dr dr V
=F + V + ,
dt dt dt t
dE dr V
= ( F + V ) + .
dt dt t
Since F is conservative
F = V ,
dE V
= .
dt t
Now if the potential energy V is independent of time t then
dE
= 0.
dt
This implies that E is conserved.
Theorem 4 : If the force acting on a particle is conservative then the total energy is
conserved.
Proof : If the particle is acted upon by the force F, then if it moves from position P1
to P2 . Hence the work done by the force is given by

P2
W = F .dr . . . (1)
P1
dv
where F = p = m ,
dt
Therefore,
P2 2 P
dv dr dv
W = m . dt = m .v dt
P1
dt dt P1
dt
P2
d 1
dt 2 mv
2
= dt ,
P1
P2
1
= mv 2 ,
2 P1
1 2 1 2
W = mv2 mv1 .
2 2
Thus W = T2 T1 . . . . (2)
Now, if the force F is conservative then it is derivable from a scalar potential
function V , which is a function of position only. Therefore, we have
V
F = V = , where V is the potential energy. Substituting this value in
r
equation (1) we get
P2
V
W = dr ,
P1
r
P2
= dV ,
P1
P
= (V ) P2 ,
1
W = V1 V2 . . . . (3)
From equations (2) and (3) we have
T2 T1 = V1 V2
T1 + V1 = T2 + V2 = constant
T + V = constant.
This shows that the total energy of the particle is conserved.

Aliter : The force field is conservative. This implies that
F = V , . . . (1)
where V is the potential energy. Newtons second law of motion defines the force by
F = mr . . . . (2)
Thus we have
V
mr = .
r
Multiply this equation by r , we get
V
mr r = r .
r
This we write as
d 1 2
mr + V = 0 .
dt 2
Integrating we get
1 2
mr + V = const.
2
This shows that the total energy of the particle moving in the conservative field of
force is constant.
Theorem 5 : If the external and internal forces are both conservative, then show that
the total potential energy V of the system is given by
1
V = Vi ( ) + Vij( ) ,
e int
i 2 i, j
where Vi ( e ) is the potential energy arises due to the external forces Fi ( e ) and Vij( int ) is
the internal energy arises due to internal forces Fji( int ) . Further show that the total
energy of the system is conserved.

Proof : Two types of forces viz., external and internal forces are acting on the system
of particles. To find the total energy of the system, we find the work done by all the
forces external as well as internal in moving the system from initial configuration 1
to the final configuration 2. It is given by
2
W = Fi dri ,
i 1
where
Fi = Fi ( e ) + Fji( int )
j
2 2
W = Fi dri + (e)
F( ji
int )
dri . . . (1)
i 1 i , j ,i j 1
Let Fi ( e ) be conservative, then there exists a potential Vi ( e ) such that
Vi (
e)
Fi ( ) = i Vi ( ) =
e e
ri
Vi ( )
2 2 e 2
Fi ( ) dri =
i
e
i ri
dri = dVi ( ) ,
i 1
e
1 1
2 2

i Fi dri = i Vi (e)
(e)
. . . (2)
1 1
Now consider the second term on the r. h. s. of equation (1)

2 2

i , j ,i j 1
Fji dri =
(int )
F(
i , j ,i j 1
ji
int )
dri . . . (3)
Interchanging i and j on the r. h. s. of equation (3) we get

2 2

i , j ,i j 1
Fji dri =
(int )
F(
i , j ,i j 1
ij
int )
drj
2 2
F
i , j ,i j 1
( int )
ji dri = F(
i , j ,i j 1
ji
int )
drj ( Fji(int ) = Fij( int ) ) . . . (4)

2 2
1

i , j ,i j 1
Fji dri =
(int )
Fji(int ) ( dri drj )
2 i , j ,i j 1
2 2
1
F
i , j ,i j 1
(int )
ji dri = Fji( int ) drij ,
2 i , j ,i j 1
for drij = dri drj
Now if the internal forces Fji( int ) are conservative, there exists a potential Vij( int ) such
that
( int ) ( int ) Vij( int )

Fji = jiV ji = ,
rij
where ji is the gradient with respect to rji . Thus the above equation becomes
2 2 (int )
1 V
F
i , j ,i j 1
( int )
ji dri = ij drij ,
2 i , j ,i j 1 rij
2 2
1
F
i , j ,i j 1
(int )
ji dri = dVij(int ) ,
2 i , j ,i j 1
2
2
1
F ( int )
ji dri = Vij(int ) . . . . (5)
i , j ,i j 1 2 i, j 1
Substituting the values from equations (2) and (5) we get
2
W = [V ]1 = V1 V2 , . . . (6)
where
1
V = Vi ( ) + Vij( )
e int
. . . (7)
i 2 i, j
represents the total potential energy of the system of particles. Similarly the total
work done by the force on the system in terms of kinetic energy is given by

2
W = Fi dri
i 1
2
d dr
W = ( mi vi ) i dt
i 1
dt dt
2
dvi
= mi vi dt ,
i 1
dt
2
d 1 2
W = mi vi dt
i 1
dt 2
2
1
W = d mi vi2
i 1 2
2
1
W = mi vi2 = T2 T1 . . . (8)
i 2 1
T1 + V1 = T2 + V2 .
This shows that the total energy of the system is conserved.
Example 8 : Find the velocity dependent potential and hence the Lagrangian for a
particle of charge q moving in an electromagnetic field.
Solution : Consider a charge particle of charge q moving with velocity v in an
electric field E and magnetic field B . The force acting on the particle is called
Lorenz force and is given by
F = q(E + v B) , . . . (1)
where E and B satisfy the Maxwells field equations

B = 0,
B . . . (2)
E = .
t

We know the vector identity A = 0 . Thus the Maxwell equation implies that
there exists the magnetic vector potential A which is a function of co-ordinates and
velocities such that
B = A. . . . (3)
Substituting this in the second Maxwell equation we get

E +
t
( A ) = 0,
A
E + = 0,
t
A
E + = 0. . . . (4)
t
We also know the vector identity
= 0 . . . (5)
Comparing equations (4) and (5) we see that, there exists a scalar potential which
is function of co-ordinates and not involving velocities such that
A
E+ =
t
A
E = . . . (6)
t
Using equations (3) and (6) in equation (1) we get
A
F = q + v A . . . (7)
t
where we have

= i +j +k ,
x y z
A A A A
=i x + j y +k z ,
t t t t

i j k
A Ay A A Ay Ax
A = = i z + j x z +k ,
x y z y z z x x y
Ax Ay Az
Ay Ax Ax Az Ay Ax Az Ay
v A = i v y vz + j vx vz +
x y z x x y y z
A A Az Ay
+ k vx x z vy
z x y z
Therefore the x-component of the Lorentz force (7) becomes

Ax Ay Ax Ax Az
Fx = q + vy vz . . . (8)
x t x y z x
Now consider
A A Ax Az Ax A A A A A
vy y x vz = vx + v y y + vz z vx x + v y x + vz x
x y z x x x x x y z
. . . (9)
Also we have
dAx A A A A
= vx x + v y x + vz x + x ,
dt x y z t
Ax A A dA A
vx + v y x + vz x = x x . . . (10)
x y z dt t
Also

( v A ) = ( vx Ax + v y Ay + vz Az )
x x

A A A
x
( v A ) = vx x + v y y + vz z
x x x
. . . (11)
Substituting from equations (10) and (11) in equation (9) we get

A A A A dA A
v y y x vz x z = ( v A ) x + x . . . . (12)
x y z x x dt t

dA
Fx = q ( v A ) x . . . (13)
x dt
Also

vx
( v A) =
vx
( vx Ax + vy Ay + vz Az ) = Ax
As is independent of vx , therefore we write

vx
( v A ) = Ax
d dA

dt vx
( v A) = x
dt
. . . (14)

d
Fx = q ( v A ) ( v A ) . . . (15)
x dt vx
Define the generalized potential
U = q ( v A ) . . . (16)
Hence we write equation (15) as

U d U
Fx = for x = vx . . . (17)
x dt x
Hence the Lagranges equation of motion

d T T
= Fx
dt x x
becomes
d T T U d U
= +
dt x x x dt x
d (T U ) ( T U )
=0
dt x x
d L L
= 0.
dt x x
where the Lagrangian of the particle L = T U becomes
1
L= m ( x 2 + y 2 + z 2 ) q + q v A . . . . (18)
2
Example 9: Show that the Lagranges equation
d T T
= Qj
dt q j q j
can also be written in the form
T T
2 = Qj .
q j q j
Solution: The kinetic energy T is in general a function of generalized co-ordinates,

generalized velocities and time. Thus we have
T = T ( q j , q j , t ) . . . . (1)
Differentiating this w. r. t. t we get

dT T T T
=T = qk + qk + . . . . (2)
dt k qk k q
k t
Differentiating equation (2) partially w. r. t. q j we get
T 2T T j 2T 2T
= qk + k + qk +
q j q
k q qk k q j qk q j t
j k

T 2T T 2T 2T
= qk + + qk + . . . . (3)
q j k q
j qk q j k q j qk q j t
Also we find the expression
d T 2T 2T 2T
= qk + qk + . . . . (4)
dt q j k qk q j k q
k q j t q j
T d T T
= . . . . (5)
q j dt q j q j
But it is given that
d T T
= + Qj .
dt q j q j
Consequently equation (5) becomes
T T T
+ Qj = .
q j q j q
j
T T
2 = Qj .
q j q j
Example10: A particle of mass M moves on a plane in the field of force given by

F = ir kr cos , where k is constant and ir is the radial unit vector. Show that
angular momentum of the particle about the origin is conserved and obtain the
differential equation of the orbit of the particle.
Solution: Let (x, y) and ( r , ) be the Cartesian and polar co-ordinates of a particle
of mass M moving on a plane under the action of the given field of force
F = ir kr cos , . . . (1)
Since the force is explicitly given, hence the Lagranges equation motion
corresponding to the generalized coordinates r and are given by
d T T
= Q , . . . (2)
dt
d T T
and = Qr , . . . (3)
dt r r
where T is the kinetic energy of the particle and is given by
1 1
T=
2 2
( )
M ( x 2 + y 2 ) = M r 2 + r 2 2 , . . . (4)
The components of generalized force along the radial direction and in the direction of
are given by
Qr = ir kr cos ,
Q = 0.
Hence equations (1) and (2) become
d
dt
(
Mr 2 = 0, )
Mr 2 = const.
and Mr Mr 2 + kr cos = 0 .

This is the equation of motion of the orbit of the particle.
Example 11: Show that the Lagranges equation of motion can also be written as
L d L
L q j =0,
t dt q j
Solution: A Lagrangian of a particle is

L = L ( q j , q j , t ) ,
Differentiating this w. r. t. we obtain

dL L L L
= q j + qj + . . . (1)
dt j q j j q
j t
Consider the expression

d L d L L
q j = q j + qj . . . (2)
dt j q j j dt q j j q j
Subtracting equation (2) from (1) we get
d L L L d L
L q j = + q j . . . . (3)
dt q j t q j dt q j
j j

But from Lagranges equation we have,
d L L
= 0. . . . (4)
dt q j q j
Consequently, equation (3) becomes
d L L
L q j =0.
dt j q j t
This is the required form.
Example 12: A particle of mass m moves in a plane under the action of a

conservative force F with components Fx = k 2 ( 2 x + y ) , Fy = k 2 ( x + 2 y ) , k is a
constant. Find the total energy of the motion, the Lagrangian, and the equations of
motion of the particle.
Solution: A particle is moving in a plane. Let (x, y) be the co-ordinates of the
particle at any instant t. If T and V are the kinetic and potential energies of the
particle then we have
1
T= m ( x 2 + y 2 ) , . . . (1)
2
and V = V ( x, y ) . . . . (2)
Since the force is given by

F = V ,
V V V
iFx + jFy = i +j +k
x y z
V V
k 2 ( 2 x + y ) i k 2 ( x + 2 y ) j = i +j ,
x y
V
= k 2 ( 2x + y ) ,
x
V
= k 2 ( x + 2 y).
y
We write
V V
dV = dx + dy ,
x y
dV = k 2 ( 2 x + y ) dx + k 2 ( x + 2 y ) dy,
dV = k 2 ( 2 xdx + d ( xy ) + 2 ydy ) .
On integrating we get,
V = k 2 ( x 2 + xy + y 2 ) . . . (3)
The total energy of motion of the particle is therefore given by

E = T +V
While the Lagrangian of the motion is given by
1
L= m ( x 2 + y 2 ) k 2 ( x 2 + xy + y 2 ) . . . (4)
2
The Lagranges equations of motion corresponding to the generalized co-ordinates x
and y are respectively given by
mx + k 2 ( 2 x + y ) = 0,
my + k 2 ( x + 2 y ) = 0.
Kinetic Energy as a Homogeneous Quadratic Function of Generalized

Velocities :
Theorem 6: Find the expression for the kinetic energy as the quadratic function of
generalized velocities. Further show that

i) when the constraints are scleronomic, the kinetic energy is a homogeneous
T
function of generalized velocities and q
j
j
q j
= 2T ,
T
ii) when the constraints are rheonomic then q
j
j
q j
= 2T2 + T1 ,
where T1 , T2 have usual meaning.
Proof: Consider a system of particles of masses mi and position vectors ri . The

kinetic energy of the system is given by
1
T=
2 i
mi ri 2 , . . . (1)
where
ri = ri ( q1 , q2 , q3 ,...qn , t ) ,
ri r
ri = qk + i .
k qk t
Substituting this value in equation (1) we get
1 r r r r
T= mi
i
q j + i i qk + i
2 i j q j t k qk t
1 ri ri ri ri ri
2
T = mi . q j qk + 2 q j + ,
2 i j , k q j qk j q j t t
1 ri ri ri ri 1 ri
2
T = mi q j qk + mi q j + mi
j ,k
2 i q j q k
j
i q j t i 2 t
or T = a jk q j qk + a j q j + a . . . (2)
j ,k j
1 r r
where a jk = mi i i ,
i 2 q j qk
r r
a j = mi i i , ... (3)
i q j t

2
1 r
a = mi i
i 2 t
are definite functions of r and t and hence functions of q ' s and t. From equation (2)
we observe that the kinetic energy is a quadratic function of the generalized
velocities.
Case 1 : If the constraints are scleronomic. This implies equivalently that the
transformation equations do not contain time t explicitly, and then we have
ri
=0,
t
and consequently a and a j vanish. Therefore equation (2) reduces to
T = a jk q j qk . . . . (4)
j ,k
This shows that the kinetic energy is a homogeneous quadratic function of

generalized velocities. Now applying Eulers theorem for the homogeneous quadratic
function of generalized velocities we have
T
q
j
j
q j
= 2T . . . (5)
Case 2 : If the constraints are rheonomic then we write equation (2) in the form
T = T2 + T1 + T0 , . . . (6)
where
T2 = a jk q j qk ,
j ,k
. . . (7)
T1 = a j q j ,
j
2
1 r
and T0 = a = mi i
i 2 t
are homogeneous function of generalized velocities of degree two, one and zero
respectively.
Now we consider
T T T T
q
j
j
q j
= q j 2 + q j 1 + q j 0
j q j j q j j q j
On applying Eulers theorem for the homogeneous function to each term on the right
hand side we readily get
T
q
j
j
q j
= 2T2 + T1 . . . . (8)
This completes the proof.
Note : However, the result (8) can also be obtained by direct differentiating equation
(2) w. r. t. q j . Thus
T
= 2 a jk qk + a j .
q j k
Next multiplying this equation by q j and summing over j we get
T
q
j
j
q j
= 2 a jk q j qk + a j q j
j ,k j
T
q
j
j
q j
= 2T2 + T1 .
The result (5) can similarly be derived by direct differentiating equation (4).
Another way of proving conservation theorem for energy :
Theorem (7): If the Lagrangian does not contain time t explicitly, the total energy of
the conservative system is conserved.
Proof : Consider a conservative system, in which the forces are derivable from a
potential V which is dependent on position only. The Lagrangian of the system is
defined as
L = T V , . . . (1)
where

L = L ( q j , q j , t ) . . . (2)
satisfies the Lagranges equation
d L L
= 0. . . . (3)
dt q j q j
Differentiating equation (2) we obtain
dL L L L
= q j + qj + .
dt q j
j q j t
L
Since L does not contain time t explicitly implies = 0.
t
dL L L
= q j + qj .
dt j
q j q j
On using equation (3) we write
dL d L L
= q j + q .
dt
j dt q j j
j q
j
dL d L
= q j ,
dt dt j q j
.
d L
L q j =0
dt j q j
L
L q j = const. . . . . (4)
j q j
Since the potential energy V for the conservative system depends upon the position
co-ordinates only and does not involve generalized velocities. Hence we have
L T
= .
q j q j
The generalized momentum is defined as

T
pj = .
q j
Thus we have from equation (4) that
p q
j
j j L = const ( H ) . . . . (5)
L does not contain time t means neither the kinetic energy nor the potential energy of
the particle involves time t. In this case the transformation equations do not contain
time t. consequently the constraints are scleronomic. Hence the kinetic energy T is a
homogeneous quadratic function of generalized velocities.
T = a jk q j qk . . . (6)
j ,k
where
1 r r
a jk = mi i i ,
i 2 q j qk
Hence by Eulers formula we have

T
q
j
j
q j
= 2T . . . (7)
Hence from equation (4) we have

2T L = H
2T T + V = H
T +V = H ,
E = H (Const )
This proves the total energy E is conserved for conservative system.
Theorem (8): Show that non-conservation of total energy is directly associated with
the existence of non-conservative forces even if the transformation equation does not
contain time t.
Proof: We know the Lagranges equations of motion for a system in which
conservative forces Fi and non-conservative forces Fi ( d ) are present are given by
d L L
= Q (j ) ,
d
j=1, 2, 3, , n. . . . (1)
dt q j q j
where the Lagrangian L contains the potential of the conservative forces and the
forces which are not arising from potential V are represented by Q(j d ) .
Since L = L ( q j , q j , t )
dL L L L
= q j + qj + . . . . (2)
dt q j
j q j t
L d L (d )
From equation (1) we have = Q j ,
q j dt q j
Therefore
dL d L L L
= q j + qj Q (j ) q j + ,
d
dt
j dt q j q
j t
j j
d L L
= q j Q(j ) q j +
d

,
j dt q j j t
dL d L L
q j Q (j ) q j +
d
= . . . (3)
dt dt j q j j t
Since L contains the potential of the conservative forces

L T
This implies that =
q j q j
dL d T L
= q j Q(j ) q j +
d
. . . . (4)
dt dt j q j j t
where T here is a quadratic function of generalized velocities and hence in this case
we have
T
q
j
j
q j
= 2T . . . . (5)

dL dT L
Q (j ) q j +
d
=2 .
dt dt j t
Hence
dE L
= Q (j ) q j .
d
. . . (6)
dt j t
If the transformation equations do not contain time t explicitly, then the kinetic
T
energy does not contain time t. This implies that = 0 . Also Lagrangian contains
t
the potential of conservative forces, we have therefore V=V ( q j ) and hence
V
= 0.
t
L
Consequently, we have = 0. Hence equation (6) becomes
t
dE
= Q (j ) q j .
d
...(7)
dt j
This shows that the non-conservation of total energy is directly associated with the
existence of non-conservative forces Q(j d ) . However, if the system is conservative
and the transformation equations do not contain time t then the total energy is
conserved.
Example 13: Show that the new Lagrangian L defined by

df ( q j , t )
L = L + , j = 1, 2,..., n
dt
satisfies Lagranges equation of motion, where f is an arbitrary differentiable
functions of q j and t , and L is a Lagrangian for a system of n degrees of freedom.
Solution : Given that

df ( q j , t )
L = L + , . . . (1)
dt
where L satisfies
d L L
= 0, . . . (2)
dt q j q j
d L L
We prove that = 0.
dt q j q j
Since f = f (q j ,t ) ,
Therefore we have
df f f
= qk + . . . . (3)
dt k qk t
Differentiating this partially w. r. t. q j we get
df 2 f 2 f
= q
k + . . . . (4)
q j dt k qk q j t q j
Also from equation (3) we have

df f
= .
q j dt q j
Differentiating this w. r. t. t we get
d df 2 f 2 f
= qk + . . . . (5)
dt q j dt k qk q j t q j
Subtracting equation (4) from (5) we get
d df df
= 0,
dt q j dt q j dt
.
d L L
i.e., =0
dt q j q j
This proves that the Lagrangian of the system is not unique.

Example 14 : Deduce the principle of energy from the Lagranges equation of
motion.
Solution : We know the Lagranges equations of motion are given by
d L L
= 0, . . . (1)
dt q j q j
where L = T V is a Lagrangian and V = V ( q j ) , T = T ( q j , q j )
d T T V
= , . . . (2)
dt q j q j q j
We also know
T
q
j
j
q j
= 2T . . . . (3)
Also we obtain
dT T T
dt
= q q + q q .
j
j
j
j . . . (4)
j j
Multiply equation (2) by q j and summing over j we obtain
d T T T V
q j qj q j = q j . . . . (5)
dt j q j j q j j q j j q j
On using (3) and (4) we write equation (5) as

d (2T ) dT dV
= ,
dt dt dt
d
(T + V ) = 0,
dt
T + V = const.
This implies that total energy is conserved.

Unit 3: Lagranges Equations for Non-holonomic Constraints:
Introduction:
We have seen that the constraints, which are not expressible in the form of
equations are called non-holonomic constraints. We have also seen that this is not the
only way to describe the non-holonomic system. A system is also said to be non-
holonomic, if it corresponds to non-integrable differential equations of constraints.
Such constraints can not be expressed in the form of equation of the type
f l ( q j , t ) = 0, l = 1, 2, 3,..., m. . . . (1)
Hence such constraints are called non-holonomic constraints. Obviously, holonomic

system has integrable differential equations of constraints expressible in the form of
equation.
Consider non-integrable differential constraints of the type
n
a
k =1
lk dqk + alt dt = 0 , . . . (2)
where alk and alt are functions of q j and t . Constraints of this type will be
holonomic only if, an integrating factor can be found that turns it in to an exact
differential, and hence the constraints can be reduced to the form of equations.
However, neither equations (2) can be integrated nor one can find an
integrating factor that will turn either of the equations in to perfect differentials.
Hence the constraints cannot be reduced to the form (1). Hence the constraints of the
type (2) are therefore non-holonomic.
Note also that non-integrable differential constraints of the type (2) are not
the only type of non-holonomic constraints. The non-holonomic constraint
conditions may involve higher order derivatives or may appear in the form of
inequalities.
There is no general way of attacking non-holonomic problems. However, the
constraints are not integrable, the differential equations of the constraint can be
introduced in to the problem along with the differential equations of motion and the
dependent equations are eliminated by the method of Lagranges multipliers. The
method is illustrated in the following theorem.
Theorem 9: Explain the method of Lagranges undetermined multipliers to construct

equations of motion of the system with non-holonomic constraints.
Proof: Consider a conservative non-holonomic system, where the equations of the
non-holonomic constraints are given by
n
a
k =1
lk dqk + alt dt = 0 , . . . (1)
where l = 1, 2, 3, ., m represents the number of constraints, and alk , alt are
functions of q j and t .
Since the constraints are non-holonomic, hence the equations expressing the
constraints (1) cannot be used to eliminate the dependent co-ordinates and hence all
the generalized co-ordinates are not independent, but are related by constraint
relations.
In the variational (Hamiltons) principle, the time for each path is held fixed
( t = 0 ). Hence the virtual displacement qk must satisfy the following equations of
constraints.
n
a q
k =1
lk k = 0, l = 1, 2, 3,..., m . . . (2)
We can use these m-equations (2) to eliminate the dependent virtual

displacement and reduce the number of virtual displacement to n-m independent one
by the method of Lagranges multipliers. Hence we multiply equations (2) by
1 , 2 , 3 , , m respectively and summing over l and integrating it between the
limits t0 to t1 we get
t1 m n
a q dt = 0
t0 l =1 k =1
l lk k . . . (3)

Hamiltons principle is assumed to hold for non-holonomic system, (see chapter 3)
we therefore have
t1
Ldt = 0 .
t0
t1 n L d L
q qk dt = 0 .
dt qk
. . . (4)
t0 k =1 k

t1
n L d L m
+
k =1 qk dt qk l =1 l alk qk dt = 0 . . . (5)
t0
Note all the virtual displacement qk , k = 1, 2,..., n are not independent but connected
by m equations (2). Now to eliminate the extra dependent virtual displacements we
choose the multipliers 1 , 2 , 3 , , m such that the coefficients of m-dependent
virtual displacements, in equation (5) are zero. i.e.,
L d L m
+ l alk = 0, for k = n ( m 1) ,..., ( n 1) , n . . . . (6)
qk dt qk l =1
Hence from equation (6) we have
t1
n m L d L m

t0

q

dt qk
l lk k dt = 0 ,
+
l =1
a q

. . . (7)
k = 1 k
where q1 , q2 , q3 ,..., qn m are all independent. Hence it follows that
L d L m
+ l alk = 0, for k = 1, 2,..., n m. . . . (8)
qk dt qk l =1
Combining equations (6) and (8) we have finally the complete set of Lagranges
equations of motion for non-holonomic system
d L L m
= l alk , k = 1, 2,..., n m,..., n . . . . (9)

dt qk qk l =1

Remarks:
1. The n-equations in (9) together with m-equations of constraints (1) are
sufficient to determine (n + m) unknowns viz., the n-generalized co-ordinates
q j and m Lagranges multipliers l .
2. Lagranges multiplier method can also be used for holonomic constraints,

when it is inconvenient to reduce all the q ' s to independent co-ordinates, and
then obtain the forces of constraints.
Worked Examples
Example 15: Use Lagranges undetermined multipliers to construct the equation of

motion of simple pendulum and obtain the force of constraint.
Solution : Consider a simple pendulum of mass m and of constant length l . Let
P (x, y) be the position co-ordinates of the pendulum. Then the equation of the
constraint is
x2 + y2 = l 2 . . . . (1)
O x This shows that x, y are not the generalized co-
ordinates. If ( r , ) are the polar co-ordinates of

l
the pendulum, then the equation of constraint is
P(x, y) (say)
f1 r l = 0 . . . . (2)
y
If this constraint is not used to eliminate the dependent variable r, then r ,

are the generalized co-ordinates. Hence the kinetic energy and potential energy of the
pendulum are respectively given by
1
T= m ( r 2 + r 2 2 ) ,
2
V = mgr cos .
The Lagrangian of the pendulum L = T V becomes
1
L=
2
( )
m r 2 + r 2 2 + mgr cos . . . . (3)
Differentiating the equation of the constraint, we get

dr = 0 .
Comparing this with the standard equation
a1r dr + a1 d = 0, (viz., alk dqk = 0 l = 1, k = 1, 2 )
k
we get
a1r = 1, a1 = 0 . . . . (4)
The Lagranges equations of motion viz.,
d L L m
= l alk , k = 1, 2,..., n
dt qk qk l =1
d L L
become = 1a1 ,
dt
d L L
and = 1a1r .
dt r r
These equations after solving become
g
+ sin = 0 , . . . (5)
r
ml + mg cos = 1 , . . . (6)
where 1 is the force of constraint, in this case it is the tension in the string. Equation
(5) determines the motion of the pendulum under the constraint force given in (6).
Example 16 : Use Lagranges undetermined multipliers to construct the equations of
motion of spherical pendulum.
Solution : Let a particle of mass m move on a frictionless surface of radius r under
the action of gravity. Let P (x, y, z) be the position co-ordinates of the pendulum. If

( r , , ) are the spherical polar co-ordinates of the pendulum, then we have the
relations
x = r sin cos ,
y
y = r sin sin , . . . (1)
z = r cos ,
P(x, y, z) = (r, , )
r where x 2 + y 2 + z 2 = r 2 .
z = r cos
This shows that x, y, z are not the generalized
x co-ordinates. The kinetic and potential energies

of the spherical pendulum are given by
respectively
z 1
T=
2
( )
m r 2 + r 22 + r 2 sin 2 2 , . . . (2)
V = mgr cos . . . . (3)

Hence the Lagrangian of the system becomes
1
L=
2
( )
m r 2 + r 2 2 + r 2 sin 2 2 mgr cos . . . . (4)
The equation of the constraint on the motion of the particle moving on the sphere is
f1 r l = 0 . . . . (5)
If this constraint is not used to eliminate the dependent variable r, then the
generalized co-ordinates are ( r , , ) . Differentiating equation (5) we get
dr = 0
Comparing this with the standard equation
a1r dr + a1 d = 0, (viz., alk dqk = 0 l = 1, k = 1, 2 ,3).
k
we get
a1r = 1, a1 = 0, a1 = 0 . . . . (6)
In this case the Lagranges equations of motion viz.,

d L L m

dt qk
= l alk , k = 1, 2,..., n
qk l =1
d L L
become = 1a1r , . . . (7)
dt r r
d L L
= 1a1 , . . . (8)
dt
d L L
and = 1a1 . . . . (9)
dt
Consequently, these equations reduce to
g
ml 2 + sin 2 2 cos = 1 . . . . (10)
l
This equation determines the constraint force. Similarly, from equations (8) and (9)
we obtain
g
sin cos 2 sin = 0 , . . . (11)
l
and sin 2 2 = p (const.) . . . (12)
Eliminating between (11) and (12) we get
p 2 g
3 cos sin = 0 . . . . (13)
sin l
Equations (10) and (13) determine the motion of the spherical pendulum.
Example 17 : A particle is constrained to move on the plane curve xy = c , where c
is a constant, under gravity. Obtain the Lagrangian and hence the equation of motion.
Solution : Given that a particle is constrained to move on the plane curve
xy = c , . . . (1)
The kinetic energy of the particle is given by
1
T= m ( x 2 + y 2 ) . . . (2)
2

The potential energy is given by
V = mgy, y is vertical . . . (3)
We see that x and y are not linearly independent as they are related by the equation
of constraint (1) and hence they are not the generalized co-ordinates. However, we
c c
eliminate the variable y by putting y = and hence y = 2 x in equations (2) and
x x
(3), we get
1 c2
T= m x 2 1 + 4 ,
2 x
c
V = mg
x
Here x is the generalized co-ordinate. Hence the Lagrangian of the particle becomes
1 2 c 2 mgc
L = m x 1 + 4 . . . . (4)
2 x x
The Lagranges equation of motion

d L L
=0 ,
dt x x
becomes
c2 c 2 m 2 mgc
mx 1 + 4 2 5 x 2 = 0 .
. . . (5)
x x x
Example 18 : A particle is constrained to move on the surface of a cylinder of fixed

radius. Obtain the Lagranges equation of motion.
Solution : The surface of the cylinder is characterized by the parametric equations
given by
x = r cos , y = r sin , z = z . . . (1)
However, x, y, z are not the generalized co-ordinates as x and y are related by the
equation of constraint x 2 + y 2 = r 2 , r is a constant radius of the circle. Hence the

generalized co-ordinates are and z. In terms of these generalized co-ordinates the
kinetic and potential energies become
1
(
T = m r 2 2 + z 2 ,
2
)
V = mgz.
Hence the Lagrangian is given by
1
L=
2
( )
m r 2 2 + z 2 mgz . . . . (2)
d L L
Therefore the Lagranges equation =0
dt
d
yields
dt
( )
mr 2 = 0 mr 2 = const (l )
Integrating we get
l
= t + 0 , . . . (3)
mr 2
where 0 is a constant of integration. Similarly, z- Lagranges equation of motion
d L L
= 0,
dt z z
1 2
gives z = ut gt , . . . (4)
2
where z = u at t = 0 .
Example 19 : A particle of mass m is projected with initial velocity u at an angle
with the horizontal. Use Lagranges equation to describe the motion of the projectile.
Solution : Let a particle of mass m be projected from O with an initial velocity u unit
making an angle with the horizontal line referred as x-axis. Let P (x, y) be the
position of the particle at any instant t. Since x and y are independent and hence the
generalized co-ordinates. The kinetic of the projectile is given by

y
1
T= m ( x 2 + y 2 ) ,
2 .
y = u sin .
u y = u sin - gt
and the potential energy is V = mgy
.
x = u cos
mg
x
Thus the Lagrangian function of the O .
x = u cos
projectile is
1
L= m ( x 2 + y 2 ) mgy . . . . (1)
2
The x- Lagranges equation of motion and y-Lagranges equation of motion
respectively give
x = 0 and
y+g =0
. . . (2)
To find the velocity of the projectile and its path at any instant we integrate equations
(2) and using boundary conditions we readily obtain
x = u cos , y = u sin gt . . . (3)
These equations determine velocity at any time t. Integrating (3) once again and
using boundary conditions we get
1 2
x = u cos .t and y = u sin .t gt . . . (4)
2
Eliminating t between equations (4) we get
1 x2
y = x tan g 2 .. . (5)
2 u cos 2
This represents the path of the projectile and it is a parabola.
Atwoods Machine :
Example 20: Explain Atwood Machine and discuss its motion.

Solution : Atwood machine consists of two masses m1 and m2 suspended over a
frictionless pulley of radius a. Both the ends of the string
a are attached the masses m1 and m2 respectively. Let the
length of the string between m1 and m2 be l . Then we
have from fig. that PA = x and QA = l - x . The system has

l- x only one degrees of freedom and x is the only generalized
x
co-ordinate. Hence the kinetic energy of the system is given
m2 Q by
1
m2g
T= ( m1 + m2 ) x 2 . . . . (1)
2
P m1 Considering the reference level as a horizontal plane
passing through A, the potential energy of both masses is
given by
m1g V = m1 gx m2 g ( l x ) . . . . (2)

1
L= ( m1 + m2 ) x 2 + ( m1 m2 ) gx + m2 gl . . . . (3)
2
The corresponding Lagranges equation of motion gives
x=

( m1 m2 ) g . . . (4)
m1 + m2
The solution of this equation gives
1 ( m1 m2 ) 2
x= gt + x0t + y0 , . . . (5)
2 m1 + m2
where x0 , y0 are constants of integration.

Example 21: A particle of mass m moves in one dimension such that it has the
Lagrangian

m 2 x 4
L= + mx 2V ( x ) V 2 ( x ) ,
12
where V is some differentiable function of x. Find equation of motion for x (t).
Solution : Here the Lagrangian of the system is
m 2 x 4
L= + mx 2V ( x ) V 2 ( x ) , . . . (1)
12
We see from equation (1) that x is the only generalized co-ordinate. Therefore the
corresponding Lagranges equation of motion becomes
V
mx + = 0. . . . (2)
x
This equation of motion shows that the particle moves in a straight line under the
V
action of a force F = .
x
Example 22 : Let a particle be moving in a field of force given by
1 r 2 2rr
F= 1 .
r2 c2
Find the Lagrangian of motion and hence the equation of motion.

Solution: One can check that F 0 , hence the force is non-conservative;
consequently, the corresponding potential is generalized potential or velocity
dependent potential. We know the component of generalized force corresponding to
the generalized coordinate r is given by
1 r2 2rr
Qr = F = 2 1 .
r c 2
We write this force as

1 r2 r
2
Qr = 2 2 2 + 2 ,
r cr cr
2
1 r r 2r2
2
Qr = 2 + 2 2 + 2 2 2 ,
r cr cr cr

1 r 2 d 1 r 2
Qr = + + + ,
r r c 2 r dt r r c 2 r
U d U
Qr = + ,
r dt r
1 r 2
U = 1 + 2 . (1)
r c
We notice that the potential energy U is the velocity dependent potential. The
kinetic energy of the particle is given by
1 2
T= mr . ... (2)
2
Hence the Lagrangian of the particle becomes
1 2 1 r 2
L = mr 1+ 2 . ...(3)
2 r c
We see that r is the only generalized co-ordinate; hence the corresponding

Lagranges equation yields the equation of motion in the form
2 r 2 1
r m 2 + 2 2 2 = 0.
(4)
rc r c r
Example 23 : Derive the equation of motion of a particle falling vertically under the
influence of gravity, when frictional forces obtainable from dissipation function
1 2
Kv are present. Integrate the equation to obtain the velocity as a function of time.
2
mg
Show also that the maximum possible velocity for fall from rest is v = .
K
Solution : Let a particle of mass m be falling vertically under the influence of
gravity. Let z be the height of the particle at any instant t. Therefore the only
generalized co-ordinate is z. Thus the Kinetic energy and the potential energy of the
particle are given by

1 2
T= mz . . . (1)
2
and V = mgz . . . (2)
Hence the Lagrangian function becomes
1 2
L= mz + mgz . . . . (3)
2
We know the Lagranges equation of motion for a system containing the frictional
1 2
forces obtainable from a dissipation function R = Kv is given by
2
d L L R
+ = 0. . . . (4)
dt z z z
Solving this equation we get
mz + Kz mg = 0 . . . . (5)
On integrating equation (5) we get
mz + Kz mgt + c1 = 0 , . . . (6)
where c1 is a constant of integration. Using initial conditions viz., when
t = 0, z = v = 0, z = 0 c1 = 0 .
We have therefore
K
z + z = gt . . . . (7)
m
This is a linear differential equation of first order whose solution is given by
2 Kt
mg m
z= t
g + c2 e m
.
K K
2
m
As t = 0 z = 0 c2 = g ,
K
Hence
2 2 Kt
mg m m
z= t g + g e .
m
. . . (8)
K K
K
Differentiating equation (8) we obtain
Kt
mg m m
z =
g e . . . . (9)
K K
This shows that the velocity z is the function of time only. For maximum velocity
we have
dz
=0.
dt
z = 0 and is given by
Hence the maximum velocity is obtained from (5) by putting
mg
z = .
K
Example 24: Two mass points of mass m1 and m2 are connected by a string passing
through a hole in a smooth table so that m1 rests on the table surface and m2 hangs
suspended. Assuming m2 moves only in a vertical line, what are the generalized co-
ordinates for the system? Write down the Lagrangian for the system. Reduce the
problem to a single second order differential equation and obtain a first integral of
the equation.
Solution: Let the two mass points m1 and m2 be connected by a string passing
suspended. We assume that m2 moves only in a vertical line. The system is shown in
the fig.
Let l be the length of a string. Consider OX as an initial line. Let ( r , ) be
the position of the particle of mass m1 .
Om2 = l r .

y
P(r, )
m1
r

O x
l- r
m2
Thus the system is specified by two generalized co-ordinates r and . The

kinetic energy of the system is the sum of the kinetic energies of the two masses and
is given by
1 1
T=
2
( )
m1 r 2 + r 2 2 + m2 r 2 .
2
. . . (1)
Potential energy of mass m1 is zero while that of mass m2 is m2 g ( l r ) .

1 1
L=
2
( 2
)
m1 r 2 + r 2 2 + m2 r 2 + m2 g ( l r ) . . . . (2)
The Lagranges equations corresponding to the generalized co-ordinates r and

respectively reduce to
( m1 + m2 ) r m1r2 + m2 g = 0 . . . (3)
and m1r 2 = const. h ( say ) . . . (4)
These are the required equations of motion. Now eliminating between (3) and (4)
we obtain
h2
( m1 + m2 ) r 3 + m2 g = 0 . . . . (5)
m1r

This is the required single second order differential equation of motion. Now to find
the first integral of (5), multiply equation (5) by 2r and integrating it w. r. t. time t,
we get
h 2 r
( m1 + m2 ) 2rrdt
m1 r 3
dt + 2m2 g rdt
= const.
2h 2 1
( m1 + m2 ) d ( r 2 ) d 2 + 2m2 g dr = const.
m1 2r
h2
( m1 + m2 ) r2 + + 2m2 gr = const. . . . (6)
m1r 2
This is required first integral of motion which represents total energy of the particle.
Example 25 : A body of mass m is thrown up an inclined plane which is moving

horizontally with a constant velocity v. Use Lagrangian equation to find the locus of
the position of the body at any time t after the motion sets in.
Solution: Let AB be an inclined plane moving horizontally with constant velocity v .
y Therefore at some instant t the distance moved
by the plane AB is given by
OA = v t . . . (1)
B
v Let at t = 0 a body of mass m be thrown up an
m inclined plane AB. Let P be the position of the
P(x, y)
r particle at that instant t, where AP = r. If
x
O A C ( x, y ) are the co-ordinates of the particle at P,
then we have
x = OA + AP cos ,
x = vt + r cos . . . (2)

and y = r sin , (note is a fixed angle ) . . . (3)
1
T= m ( x 2 + y 2 ) .
2
We notice that x and y are related by equations (2) and (3) and hence will not be the
generalized co-ordinates. The only generalized co-ordinate is r. Hence using
equations (2) and (3) we write the expression for the kinetic energy in terms of
generalized co-ordinate r as
1
T= m ( v 2 + r 2 + 2rv
cos ) . . . . (4)
2
The potential energy of the particle is given by
V = mgr sin . . . . (5)
Hence the Lagrangian of the system is
1
L= m ( v 2 + r 2 + 2rv
cos ) mgr sin . . . . (6)
2
Hence the corresponding r- Lagranges equation reduces to the form
r = g sin .
. . . (7)
Integrating we get
r = g sin t + c1 .
At t = 0 let r = u be the initial velocity of the particle with which it is projected.
This gives c1 = u , hence
r = u g sin t . . . . (8)
Integrating once again we get
1 2
r = u.t gt sin + c2 .
2
At t = 0, r = 0 c2 = 0.
1 2
r = ut gt sin .
2
Hence the locus of the position of the particle is given by
2
1 2
r = ut gt 2 sin = ( x vt ) + y 2 .
2
. . . (9)
2
Example 26 : Set up the Lagrangian and the Lagranges equation of motion for
simple pendulum.
O x
Solution : Consider a simple pendulum of point
mass m attached to one end of an inextensible light
l
string of length l and other end is fixed at point O.
The system is shown in fig. If B ( x, y ) are the C
B(x, y)
position co-ordinates of the pendulum at any instant
t, then the equation of the constraint is given by
x2 + y2 = l 2 , . . . (1)
where x = l sin , y = l cos , is the angle made by the pendulum with the
vertical. This shows that x and y are not the generalized co-ordinates. We see that
the angle determines the position of pendulum at any given time; hence it is a
generalized co-ordinate. Hence the kinetic and potential energies of the pendulum
become
1 2
T=
2
( )
m l , V = mgl (1 cos ) . . . . (2)
Hence the Lagrangian of the motion becomes

1 2
L=
2
( )
m l mgl (1 cos ) . . . . (3)
The -Lagranges equation of motion gives

g
+ sin = 0 . . . . (4)
l

This is the second order differential equation that determines the motion of the
simple pendulum.
Example 27 : A pendulum of mass is attached to a block of mass M. The block
slides on a horizontal frictional less surface. Find the Lagrangian and equation of
motion of the pendulum. For small amplitude oscillations derive an expression for
periodic time.
y
Solution : Let a pendulum of point mass m be x1
O M x
attached to one end of the light and inextensible
string of length l and other end is attached to a l
block of mass M. The system is shown in fig. Let at
x2
any instant t the position co-ordinates of the block P(x 2 , y2)
of mass M and the pendulum of mass m be ( x1 , 0 )
and ( x2 , y2 ) respectively,
where x2 = x1 + l sin ,
y2 = l cos .
We see that the position co-ordinates of the pendulum are related by the constraint
equation; hence these are not the generalized co-ordinates. The generalized co-
ordinates in this case are x1 and . The kinetic energy of the system is the sum of the
kinetic energy of the pendulum and the kinetic energy of the block. It is given by
1 1
T=
2 2
( )
Mx12 + m x12 + l 2 2 + 2lx1 cos .
The potential energy of the pendulum is given by

V = mgl cos .
1 1
L=
2 2
( )
( M + m ) x12 + m l 22 + 2lx1 cos + mgl cos . . . . (1)

Since and x1 are the generalized co-ordinates, hence the corresponding
Lagranges equations of motion viz.,
d L L d L L
= 0 and =0,
dt dt x1 x1
respectively reduces to
ml 2 + ml cos
x1 + mgl sin = 0 , . . . (2)
( M + m ) x1 + ml = 0 . . . . (3)
However, if is small then we have sin = and cos = 1 . Consequently equation

(2) becomes
x1 g

+ + =0 . . . (4)
l l
Eliminating x1 between equations (3) and (4) we get
=
( M + m) g . . . . (5)
Ml
This is the required equation of simple harmonic motion. The periodic time T is
given by
2
T=
accel n . per unit displacement
Ml
T = 2 .
( M + m) g
Spherical Pendulum: A point mass constrained to move on the surface of a
sphere is called spherical pendulum.
Example 28 : In a spherical pendulum a particle of mass m moves on the surface of
a sphere of radius r in a gravitational field. Show that the equation of motion of the
particle may be written as
p2 cos g
2 4 3
sin = 0,
m r sin r
where p is the constant of angular momentum.
Solution : Let P (x, y, z) be the position co-ordinates of the particle moving on the
surface of a sphere of radius r. If ( r , , ) are its spherical co-ordinates, then we have
z x = r sin cos ,
y = r sin sin , . . . (1)
P(r, , ) z = r cos
It clearly shows that x, y, z are not the generalized
r
z = r cos co-ordinates, as they are related by the constraint
equations (1). The generalized co-ordinates are
y
( , ) . Hence the kinetic and potential energies of
L
particle are respectively given by
x
1 2 2
T=
2
( )
mr + sin 2 2 ,
V = mgr cos
Hence the Lagrangian function becomes
1 2 2
L=
2
(
mr + sin 2 2 mgr cos . ) . . . (2)
The two Lagranges equations of motion corresponding to the generalized co-

ordinates and reduce to
mr 2 mr 2 sin cos 2 mgr sin = 0 . . . (3)
and mr 2 sin 2 = const. = p . . . . (4)
Eliminating between equations (3) and (4) we get
p2 cos g
2 4 3
sin = 0, . . . (5)
m r sin r
where p is a constant of angular momentum.
Compound Pendulum :
A rigid body capable of oscillating in a vertical plane about a fixed horizontal
axis under the action of gravity is called a compound pendulum.

Example 29 : Set up the Lagrangian and the Lagranges equation of motion for the
compound pendulum.
Solution : Let O be a fixed point of a rigid body through which axis of rotation
passes. Let C be the center of mass, and OC = l .Let m be the mass of the pendulum
and I the moment of inertia about the axis of rotation. If is the angle of deflection
of the body then the rotational kinetic energy of the pendulum is given by
y
1 2
T= I . . . (1)
2
O
l The potential energy relative to the horizontal plane
C through O is
V = mgl cos . . . (2)
Hence the Lagrangian of compound pendulum becomes
mg
1 2
O x L= I + mgl cos . . . (3)
2
Thus the Lagranges equation of motion corresponding to the generalized co-
ordinate becomes
mgl
+ sin = 0 . . . . (4)
I
The periodic time of oscillation is given by
I
T = 2 . . . . (5)
mgl
Example 30 : Obtain the Lagrangian and equations of motion for a double

pendulum vibrating in a vertical plane.
Solution : A double pendulum moving in a plane consists of two particles of masses
m1 and m2 connected by an inextensible string. The system is suspended by another
inextensible and weightless string fastened to one of the masses as shown in the fig.
Let 1 and 2 be the deflections of the pendulum from vertical. These are the

generalized co-ordinates of the system. Let l1 and l2 be O x
1 l1
the lengths of the strings and ( x1 , y1 ) , ( x2 , y2 ) be the
(x 1, y 1)
rectangular position co-ordinates of the masses m1 and l2
2
m2 respectively at any instant t. From the fig. we have (x 2, y 2)
y
x1 = l1 sin 1 , y1 = l1 cos 1 ;
x2 = l1 sin 1 + l2 sin 2 , . . (1)
y2 = l1 cos 1 + l2 cos 2
The total kinetic energy of the system is given by
1 1
T= m1 ( x12 + y12 ) + m 2 ( x22 + y 22 ) . . . . (2)
2 2
Using equation (1) we obtain
1 1
T= m1l1212 + m2 l1212 + l2 222 + 2l1l212 cos (1 2 ) . . . . (3)
2 2
Taking the reference level as a horizontal plane through the point of suspension O,
the total potential energy of the system is given by
V = m1 gl1 cos 1 m2 g ( l1 cos 1 + l2 cos 2 ) . . . . (4)

1 1
L= ( m1 + m2 ) l1212 + m2l2222 + m2l1l2 cos (1 2 )12 +
2 2
+ m1 gl1 cos 1 + m2 g ( l1 cos 1 + l2 cos 2 ) . . . . (5)
Solving the Lagranges equations of motion corresponding to the generalized co-

ordinates 1 and 2 we obtain
( m1 + m2 ) l121 + m2l1l2 cos (1 2 ) 2 + m2l1l2 sin (1 2 ) 212 + ( m1 + m2 ) gl1 sin 1 = 0 , . . (6)
and
m2l222 + m2l1l2 cos (1 2 ) 1 m2l1l2 sin (1 2 )112 + m2 gl2 sin 2 = 0 . . . . (7)
Equations (6) and (7) describe the motion of the double pendulum.
Note : If in particular, two masses are equal, the lengths of the pendula are also
equal and 1 2 is very small, then for small angle we have sin = , cos = 1
and hence neglecting the terms involving 2 we get from equations (6) and (7) that
2l1 + l2 + 2 g1 = 0,
. . . (8)
l + l + g = 0.
2 1 2
Example 31: A particle is moving on a cycloid s = 4a sin under the action of

gravity. Obtain the Lagrangian and Lagranges equation of motion.
Solution : A particle is moving on a cycloid under the action of gravity whose
intrinsic equation is given by
s = 4a sin . . . (1)
From equation (1) we find
ds = 4a cos d ,
ds 2 = 16a 2 cos 2 d 2 . . . . (2)
Hence the kinetic energy of the particle is given by
2
1 ds
T = m
2 dt . . . (3)
T = 8a 2 m cos 2 2 .
To find the potential energy of the particle, let P (x, y) be the position of the particle
at any instant, where the Cartesian equations of the cycloid are given by
x = a ( 2 + sin 2 ) ,
. . . (4)
y = a (1 cos 2 ) .
The potential energy of the particle is therefore

V = mgy ,
V = mga (1 cos 2 ) . . . . (5)
Thus the Lagrangian of motion of the particle is

L = 8a 2 m cos 2 2 mga (1 cos 2 ) . . . (6)
We see that the system has one degree of freedom and is the only generalized co-
ordinate. Hence the Lagranges equation of motion is obtained as
g
tan 2 + tan = 0 . . . . (7)
4a
Example 32: Obtain the expression for kinetic energy of a particle constrained to
move on a horizontal xy plane which is rotating about the vertical z-axis with
angular velocity . Show that
T T
x + y = 2T2 + T1 ,
x y
1
where T2 = m ( x 2 + y 2 ) ,
2
T1 = m ( xy yx )
Show also that the Lagrangian of the particle is given by

1
m ( x y ) + ( y + x ) V ( x, y ) .
2 2
L=
2
Solution : A particle is moving on the xy -plane and the plane itself is rotating with
respect to z-axis with angular velocity . Let ( x1 , y1 , z1 ) be the co-ordinates of the
particle with respect to the fixed co-ordinate system and ( x, y, z ) the co-ordinates of
the particle with respect to rotating axes. The co-ordinates with respect to the rotating
axes are taken as the generalized co-ordinates. The transformation equations for
rotation are given by
x1 = x cos t y sin t ,
. . . (1)
y1 = x sin t + y cos t ,
z1 = z . . . (2)
Since z fixed is the constraint, therefore, the system has only two degrees of freedom
and hence only two generalized co-ordinates and that are x and y. We note here that

the transformation equations (1) are not independent of time, though the constraint
equation (2) is. Thus the kinetic energy of the particle is given by
1
T= m ( x12 + y12 ) . . . (3)
2
Differentiating equations (1) with respect to t and putting in (3) we obtain
1
T= m ( x 2 + y 2 ) + 2 ( x 2 + y 2 ) + 2 ( xy yx ) ,
2
1
m ( x y ) + ( y + x ) .
2 2
T= . . . (4)
2
We can also write this equation as
T = T2 + T1 + T0 ,
where
1
T2 = m ( x 2 + y 2 ) ,
2
T1 = m ( xy yx ) ,
1
T0 = m 2 ( x 2 + y 2 ) .
2
Differentiating (4) w. r. t. x, y we get
T
= mx my,
x
T
= my + mx.
y
This gives on solving
T T
x + y = 2T2 + T1 ,
x y
Now if V is the potential energy of the particle which is function of the generalized
co-ordinates x, y then the Lagrangian of the particle is given by
L = T V ,
1
m ( x y ) + ( y + x ) V ( x, y ) .
2 2
L= . . . (5)
2
Example 33 : The Lagrangian of a system is
m k
L=
2
( + cy 2 ) ( ax 2 + 2bxy + cy 2 ) , a, b, c
ax 2 + 2bxy
2
are arbitrary constants such that b 2 4ac 0 . Write down the equation of motion.
Examine the two cases a = 0, c = 0 and b = 0, c = a and interpret physically.
Solution: Given that
m k
L=
2
( + cy 2 ) ( ax 2 + 2bxy + cy 2 ) ,
ax 2 + 2bxy
2
. . . (1)
a, b, c are arbitrary constants. We notice that x and y are the generalized co-ordinates.
Hence the corresponding Lagranges equations of motion are
m ( ax + by) + k ( ax + by ) = 0 , . . . (2)
m ( bx + cy) + k ( bx + cy ) = 0 . . . . (3)
Case (i) If a = 0, c = 0 .
Equations (2) and (3) reduce to
k
y + y = 0,
. . . (4)
m
k
and x+ x = 0.
. . . (5)
m
Case (ii) b = 0, c = a . Putting this in equations (2) and (3) we get
k
x+ x = 0,
. . . (6)
m
k
and y + y = 0.
. . . (7)
m
We see from the equations (4), (5) and (6), (7) that in both the cases we get the same
set of equations of motion. These are the differential equations of particle performing

a linear S. H. M. The solution of these equations gives the displacement of the
k
particle with the frequency of oscillation = .
m
Unit 4: Generalized Momentum and Cyclic co-ordinates:

Introduction: We have proved some conservation theorems in the Unit 1. In this
unit we will prove that the conservation theorems are continued to be true for cyclic
generalized coordinates.
Definitions: In Newtonian mechanics the components of momentum (linear) are
defined as the derivative of kinetic energy with respect to the corresponding
components of velocity. i.e., If
1
T= m ( x 2 + y 2 + z 2 )
2
is the kinetic energy of a particle, then the components of momentum of the particle
are defined as
T T T
px = = mx , p y = = my , pz = = mz . . . . (1)
x y z
Generalized Momentum :
Consider a conservative system in which the forces are derivable from a potential
function V which is dependent on position only. In this case we have
L T
= .
q j q j
Thus the quantity

L
pj = . . . . (2)
q j
is called the generalized momentum associated with the generalized co-ordinates q j .

Note 1 : The definition of generalized momentum (2) is exactly analogous to the
usual definition of momentum (1).
Note 2 : The word generalized momentum subsumes linear momentum and
angular momentum of the particle.
e.g. To illustrate, let a particle be moving in plane polar co-ordinates ( r , ) . Then we
have its kinetic energy is given by

1
T=
2
( )
m r 2 + r 2 2 .
Hence the generalized momentum corresponding to the generalized co-ordinate r and

are respectively given by
T
pr = = mr ,
r
T
p = = mr .

We notice that pr and p represent respectively the linear momentum and angular
momentum of the particle.
Cyclic or ignorable Co-ordinates:

Co-ordinates which are absent in the Lagrangian are called cyclic or
ignorable co-ordinates, although the Lagrangian may contain the corresponding
generalized velocity q j of the particle.
Conservation Theorem for generalized momentum :

Theorem 10 : Show that the generalized momentum corresponding to a cyclic co-
ordinate is conserved.
Proof : The Lagranges equations of motion are given by
d L L
= 0, . . . (1)
dt q j q j

where L = L ( q j , q j , t ) is the Lagrangian function. If the generalized co-ordinate q j is
cyclic in L, then it must be absent in the Lagrangian. Obviously we have therefore,

L
=0. . . . (2)
q j
Thus the Lagranges equation of motion (1) becomes
d L
= 0 . . . (3)
dt q j
L
But we have pj = .
q j
This implies that

d
dt
( p j ) = 0,
p j = const.
This proves that the generalized momentum corresponding to the cyclic co-ordinate
is conserved.
Conservation Theorem for Linear momentum :
We will show that the conservation Theorems are continued to be true for
cyclic generalized co-ordinates.
Theorem 12 : If the cyclic generalized co-ordinate q j is such that dq j represents
the translation of the system, then prove that the total linear momentum is conserved.
Proof : Consider a conservative system so that the potential y
energy V is a function of generalized co-ordinates only.
P
i.e. V = V ( q j ) .
>
dqj n
ri(q j)
Hence we have Q
V ri (q j + dq j)
=0. . . . (1)
q j O x

Let P = ri ( q j ) be the initial position of the system and let it be translated to a point
Q = ri ( q j + dq j ) , so that
PQ = dq j n, ...(2)
where n is the unit vector along the direction of translation and dq j represents the
translation of the system.

We know by the first principle that
ri ri ( q j + dq j ) ri ( q j )
= lim as dq j 0,
q j dq j
ri PQ
= lim as dq j 0,
q j dq j
ri dq
= lim j n as dq j 0.
q j dq j
This gives on using (1)

ri
= n. . . . (3)
q j
Now the generalized force is given by

ri
Q j = Fi .
i q j
On using equation (3) we get

Q j = Fi n Q j = Fn, . . . (4)
i
where F is the total force acting on the system. Equation (4) implies that Q j are the
components of the total force in the direction of translation n .

Now the generalized momentum p j is defined by
T
pj = , . . . (5)
q j
where T is the kinetic energy of the system and is given by

1
T = mi ri 2 .
i 2
Thus we have
1
pj =
q j
2m r ,
i
i i
2
ri
p j = mi ri ,
i q j
ri ri r
p j = mi ri , as = i .
i q j q j q j

p j = mi ri n ,
i
p j = pi n,
i
p j = p n,
where p is the total linear momentum of the system. This equation shows that p j are
the components of total linear momentum of the system along the displacement dq j .
Since in the translation of the system, velocity is not affected and hence the kinetic
energy of the system. This means that q j will not appear in kinetic energy
expression. That is, change in the kinetic energy due to change in q j is zero.
Consequently, we have
T
=0. . . . (6)
q j
Thus from the Lagranges equation of motion on using equations (1) and (6) we
have,
d T V
+ =0
dt q j q j

V
p j = = Qj . . . . (7)
q j
Now, if the co-ordinate q j is cyclic in the Lagrangian, then
L
=0
q j
Due to equation (6), we have

V
=0
q j
Consequently, we have from equation (7)

p j = 0 p j = const. . . . (8)
This shows that corresponding to the cyclic co-ordinate q j the total linear
momentum is conserved.
Note : This can also be stated from equation (7) that if the components of total force
Q j are zero, then the total linear momentum is conserved.
Theorem 13 : If the cyclic generalized co-ordinate q j is such that dq j represents
the rotation of the system of particles around some axis n , then prove that the total
angular momentum is conserved along n .
>
Proof : Consider a conservative system so that the potential n
energy V is a function of generalized co-ordinates only.

M
dqj
V = V (q j ) . Q
P
dr i
ri(qj)
Hence we have
ri (q j + dqj)
V
=0. . . . (1)
q j dqj
Let the system be rotated through an angle dq j around a unit

O
vector n . This gives the rotation of the vector

OP = ri ( q j ) to OQ = ri ( q j + dq j )
The magnitude of change in position vector ri ( q j ) due to rotation is given by
dri = MPdq j ,
= OP sin dq j ,
dri = ri ( q j ) sin dq j
dri
= ri sin . . . . (2)
dq j
Since ri = ri ( q1 , q2 , q3 ,..., qn , t ) , we write therefore from equation (2)
ri
= n ri . . . . (3)
q j
ri
This shows that is perpendicular to both n and ri .
q j
The generalized force is given by

ri
Q j = Fi .
i q j

Q j = Fi ( n ri )
i
Q j = n ( ri Fi ) . . . (4)
i
= n N i ,
i
where N i = ri Fi is the torque on the i th particle. If N = N i is the total torque

i
acting on the system, then equation (4) shows that Q j are the components of the total
torque along the axis of rotation. Now the generalized momentum p j is defined by

T
pj = . . . . (5)
q j
where T is the kinetic energy of the system and is given by

1
T = mi ri 2 ,
i 2
Thus we have
1
pj =
q j
2m r ,
i
i i
2
ri
p j = mi ri ,
i q j
ri ri r
p j = mi ri , as = i .
i q j q j q j

p j = mi ri ( n ri ) ,
i
p j = pi ( n ri ) ,
i
p j = n ( ri pi ),
i
p j = n ( ri pi )
i
p j = n Li
i
where L = Li is the total angular momentum of the system.

i
Thus we have
p j = nL
. . . . (6)
This equation shows that p j are the components of total angular momentum of the
system along the axis of rotation. Since the rotation of the system does not change
the magnitude of the velocity and hence the kinetic energy of the system. This means

that T does not depend on positions q j . That is, change in the kinetic energy due to
change in q j is zero. Consequently, we have
T
=0. . . . (7)
q j
Thus from the Lagranges equation of motion on using equations (1) and (7) we
have,
d T V
+ =0
dt q j q j
V
p j = = Qj . . . . (8)
q j
Now, if the co-ordinate q j is cyclic in the Lagrangian, then
L
=0.
q j
Due to equation (7), this gives

V
=0.
q j
Consequently, we have from equation (8)

p j = 0 p j = const. . . . (9)
This shows that corresponding to the cyclic co-ordinate q j the total angular
momentum is conserved.
Note : From equation (8) the Theorem can also be stated as, if the applied torque is
zero then the total angular momentum is conserved.

Exercise:
1. Show that
d T
i) L = p j q j , ii) 2 = q j , iii) T + V = 0
dt j p j
for conservative scleronomic systems.

2. Derive the Newtons equation of motion from the Lagranges equation of
motion for a particle moving under the action of the force F.
Hint : The force is explicitly given, use Lagranges equation of motion
d T T
= Qj ,
dt q j q j
where T is the kinetic energy of the particle and is given by
1
T= m ( x 2 + y 2 + z 2 ) .
2

3. A particle is moving on a cycloid s = 4a sin under the action of gravity.
2
Obtain the Lagrangian and the equation of motion.

L = 2a 2 m cos 2 2 mga (1 cos ) ,
2
Ans:
1 g
tan 2 + tan = 0
2 2 2a 2
4. Show that the force F defined by
F = ( y 2 z 3 6 xz 2 ) i + 2 xyz 3 j + ( 3 xy 2 z 2 6 x 2 z ) k
is conservative and hence find the total energy of the particle.

Ans : For conservative force
1
F = 0, E = m ( x 2 + y 2 + z 2 ) + 3 x 2 z 2 xy 2 z 3 .
2
5. Show that Newtons equation of motion is the necessary condition for the
action to have the stationary value.

Show that the two Lagrangians L1 = ( q + q ) , L2 = ( q 2 + q 2 ) are equivalent.
2
6.
Ans : Both the Lagrangians produce the same equation of motion q q = 0 .
7. For a mechanical system the generalized co-ordinates appear separately in the
kinetic energy and the potential energy such that
T = f i ( qi ) qi2 , V = Vi ( qi ) .
Show that the Lagranges equations reduce to

2 f i qi + f i qi2 + Vi = 0, i = 1, 2,..., n .
8. The length of a simple pendulum changes with time such that l = a + bt ,

where a and b are constants. Find the Lagrangian and the equation of motion.
Ans : Equation of motion ( a + bt ) + 2b + g sin = 0.
9. Find the Lagrangian and the equation of motion of a particle of mass m

moving on the surface characterized by
x = r cos , y = r sin , z = r cot .
1
Ans : L=
2
( )
m r 2 cos ec 2 + r 22 mgr cot .
r r sin 2 2 + g cos sin = 0 .

And equation of motion is
10. Find the Lagrangian and the equation of motion of a particle moving on the
surface obtained by revolving the line x = z about z - axis.
Hint : Surface of revolution is a cone x 2 + y 2 = z 2
1
Ans : Lagrangian of the particle L =
2
( )
m 2r 2 + r 22 mgr .
1 1
r r2 + g = 0 .
Equation of motion
2 2
11. Describe the motion of a particle of mass m moving near the surface of the
earth under the earths gravitational field by Lagranges procedure.

1
Ans.: L = m ( x 2 + y 2 + z 2 ) mgz ,
2
x = 0,
and equations of motion are y = 0,
z = g .
12. A particle of mass m can move in a frictionless thin circular wire of radius r.
If the wire rotates with an angular velocity about a vertical diameter,
deduce the differential equation of motion of the particle.
g
Ans : Equation of motion of the particle : 2 sin cos sin = 0.
r

CHAPTER - II
VARIATIONAL PRINCIPLES
Unit 1: Euler-Lagrangess Differential Equations:

Introduction:
We have seen that co-ordinates are the tools in the hands of a mathematician.
With the help of these co-ordinates the motion of a particle and also the path
followed by the particle can be discussed. The piece wise information of the
path y = f ( x ) , whether it is minimum or maximum at a point can be obtained from
differential calculus by putting y = 0 . The function is either maximum or minimum at

the point depends upon the value of second derivative of the function at that point.
The function is maximum at a point if its second derivative is negative at the point,
and is minimum at the point if its second derivative is positive at that point.
However, if we want to know the information about the whole path, we use
integral calculus. i.e., the techniques of calculus of variation and are called
variational principles. Thus the calculus of variation has its origin in the
generalization of the elementary theory of maxima and minima of function of a
single variable or more variables. The history of calculus of variations can be traced
back to the year 1696, when John Bernoulli advanced the problem of the
brachistochrone. In this problem one has to find the curve connecting two given
points A and B that do not lie on a vertical line, such that a particle sliding down this
curve under gravity from A reaches point B in the shortest time.
Apart from the problem of brachistochrone, there are three other problems
exerted great influence on the development of the subject and are:
1. the problem of geodesic,
2. the problem of minimum surface of revolution and
3. the isoperimetric problem.
Thus in calculus of variation we consider the motion of a particle or system
of particles along a curve y = f(x) joining two points P ( x1 , y1 ) and Q ( x2 , y2 ) . The
infinitesimal distance between two points on the curve is given by

1
P(x2 , y2 )
ds = ( dx 2 + dy 2 ) 2 .
Hence the total distance between two point P and Q

Q(x2 , y 2) ds along the curve is given by
x2 1
dy
I ( y ( x)) = (1 + y ) dx,
2 2
y =
x1
dx
In general the integrand is a function of the independent variable x, the

dependent variable y and its derivative y . Thus the most general form of the integral
is given by
x2
I ( y ( x)) = f ( x, y, y)dx.
x1
. . . (1)
This integral may represents the total path between two given points, the surface
area of revolution of a curve, the time for quickest decent etc. depending upon the
situation of the problem. The functional I in general depends upon the starting point
( x1 , y1 ) , the end point ( x2 , y2 ) and the curve between two points. The question is
what function y is of x so that the functional I ( y ( x ) ) has stationary value. Thus in
this chapter we first find the condition to be satisfied by y(x) such that the functional
I ( y ( x ) ) defined in (1) must have extremum value. The fascinating principle in
calculus of variation paves the way to find the curve of extreme distance between
two points. Its object is to extremize the values of the functional. This is one of the
most fundamental and beautiful principles in applied mathematics. Because from this
principle one can determine the
(a) Newtons equations of motion,
(b) Lagranges equations of motion,
(c) Hamiltons equations of motion,
(d) Schrdingers equations of motion,
(e) Einsteins field equations for gravitation,
(f) Hoyle-Narlikars equations for gravitation and so on and so forth by
slightly modifying the integrand.
Note : A functional means a quantity whose values are determined by one or several
functions. i.e., domain of a functional is a set of all admissible functions.
e.g. The length of the path l between two points is a function of curves y(x),
which it self is a function of x. Such functions are called functional.
Basic Lemma :
If x1 and x2 ( > x1 ) are fixed constants and G ( x ) is a particular continuous
x2
function for x1 x x2 and if G ( x ) ( x ) dx = 0

x1
for every choice of continuous
differentiable function ( x ) such that ( x1 ) = 0 = ( x2 ) , then G ( x ) = 0 identically
in x1 x x2 .
Proof : Let the lemma be not true. Let us assume that there is a particular value x of
x in the interval such that G ( x ) 0 . Let us assume that G ( x ) > 0.
x
[ ( ) ]
x1 x1 x2 x2
Since G ( x ) is continuous function in x1 x x2 and in particular it is
continuous at x = x . Hence there must exist an interval surrounding x say

x1 x x2 in which G ( x ) > 0 everywhere.

x2
Let us now see whether the integral G ( x ) ( x ) dx = 0

x1
permissible choice
of ( x ) .
We choose ( x ) such that
( x) = 0 for x1 x x1
2 2
= ( x x1 ) ( x x2 ) for x1 x x2 . . . (1)
= 0 for x2 x x2
For this choice of ( x ) which also satisfies
( x1 ) = ( x2 ) = 0 ,
x2
the integral G ( x ) ( x ) dx becomes

x1
x2 x1 x2 x2
G ( x ) ( x ) dx = G ( x ) ( x ) dx + G ( x ) ( x ) dx + G ( x ) ( x ) dx
x1 x1 x1 x2
x2 x2
G ( x ) ( x ) dx = ( x x ) ( x x ) G ( x ) dx
2 2
1 2 . . . (2)
x1 x1
Since
G ( x ) > 0 in x1 x x2 ,
R. H. S. of equation (2) is definitely positive.

x2
( x )G ( x ) dx > 0,
x1
This is contradiction to the hypothesis

x2
G ( x ) ( x ) dx = 0 .
x1
If G ( x ) < 0 ,

we obtain the similar contradiction. This contradiction arises because of our
assumption that G ( x ) 0 for x in x1 x x2 .
This implies that G ( x ) = 0 identically in x1 x x2 .
This completes the proof.

Theorem 1 : Find the Euler- Lagrange differential equation satisfied by twice
differentiable function y(x) which extremizes the functional
x2
I ( y ( x )) = f ( x, y, y)dx
x1
where y is prescribed at the end points.

Proof: Let P ( x1 , y1 ) and Q ( x2 , y2 ) be two fixed y (
x)
0 )+
y (x,
points in xy plane. The points P and Q can be joined )= Q(x 2, y 2)
y ( x, 0 )
c: x,
by infinitely many curves. Accordingly the value of y(
y=
c:
the integral I will be different for different paths. We
P(x1 , y 1)
shall look for a curve along which the functional I x
O
has an extremum value. Let c be a curve between P
and Q whose equation is given by y = y ( x,0 ) .
Let also the value of the functional along the curve c be extremum and is given by
x2
I ( y ( x )) = f ( x, y, y)dx
x1
. . . (1)
We can label all possible paths starting from P and ending at Q by the family of
equations
y ( x, ) = y ( x, 0 ) + ( x ) , . . . (2)
where is a parameter and ( x ) is any differentiable function of x.

For different values of we get different curves. Accordingly the value of the
integral I will be different for different paths. Since y is prescribed at the end points,
this implies that there is no variation in y at the end points. i.e., all the curves of the
family must be identical at fixed points P and Q.
( x1 ) = 0 = ( x2 ) . . . (3)
Conversely, the condition (3) ensures us that the curves of the family that all pass
through the points P and Q. Let the value of the functional along the neighboring
curve be given by
x2
I ( y ( x, ) ) = f ( x, y ( x, ) , y ( x, ) ) dx
x1
. . . (4)
I
From differential calculus, we know the integral I is extremum if = 0,
= 0
since for = 0 the neighboring curve coincides with the curve which gives
extremum values of I .
x2
I f f
Thus = 0,
= 0

x1
( x ) + ( x ) dx = 0 .
y y
Integrating the second integration by parts, we get
x2 x2 x
f f 2
d f
x y ( x ) dx +
y
( x ) ( x ) dx = 0
x1 x1 dx y
. . . (5)
As y is prescribed at the end points, hence on using equations (3) we obtain

x2
f d f
y dx y ( x ) dx = 0 .
x1
By using the basic lemma of calculus of variation we get

f d f
= 0. . . . (6)
y dx y
This is required Euler- Lagrange differential equation to be satisfied by y(x) for
which the functional I has extremum value.
Important Note :
If however, y is not prescribed at the end points then there is a difference in y
x2
even at the end points and hence ( ( x ) ) x 0 . As the value of the functional I is
1
taken only on the extremal between two points and hence we must have the Euler-
Lagrange equation is true. Consequently, in this case we must have from equation (5)
that
x2
f f f
= 0 = 0 and = 0. . . . (7)
y x1 y x1 y x2
We will prove this result a little latter in Theorem No. 2.
Aliter : (Proof of the above Theorem (1)):
Let P ( x1 , y1 ) and Q ( x2 , y2 ) be two fixed points in xy plane. Let c be the
curve between P and Q whose equation is given by y = y(x).Let the extremum value
of the functional along the curve c be given by
x2
I ( y ( x ) ) = f ( x, y, y )dx . . . . (1)
x1
To find the condition to be satisfied by y(x), let

y
the curve c be slightly deformed from the original
position such that any point y on the curve c is y
y+
y Q(x2 , y2)
displaced to y + y , where y is the variation in 0)
= y(x,
c : y
the path for an arbitrary choice of , at any point
P(x1 , y1 )
except at the end points, as y prescribed there.
x
O
Mathematically this means that
y ( x1 ) = y1 , y ( x2 ) = y2
x2
( y ) x =0. . . . (2)
1

Thus the value of the functional along the varied path is given by
x2
I = f ( x, y + y, y + y)dx .
x1
. . . (3)
Hence the change in the value of the functional due to change in the path is given by
x2
I I = f ( x, y + y, y + y ) f ( x, y, y ) dx ,
x1
x2
Let I I = I = f ( x, y + y, y + y ) f ( x, y, y ) dx . . . . (4)
x1
We recall the Taylors series expansion for the function of two variables
f f
f ( x, y + y , y + y ) = f ( x , y , y ) + y + y + ...
y y
Since y is very small, therefore by neglecting the higher order terms in y and
y we have
f f
f ( x, y + y , y + y ) f ( x, y , y ) = y + y .
y y
Substituting this in the equation (4) we get
x2
f f
I = y + y dx
x
y1
y
dy d
We know = y,
dx dx
hence we have
x2
f f d
I = y+ ( y ) dx .
x1
y y dx
Integrating the second integral by parts we get
x2 x2 x
f f 2
d f
I = y dx + y y dx .
x1
y y x1 x1 dx y

x2
f d f
I = y dx . . . . (5)
x1
y dx y
If I is extremum along the curve y = y(x) then change in I is zero. ( i.e., I = 0 ) .
This resembles very closely with a similar condition of extremum of a function in

differential calculus.
x2
f d f
y dx y y dx = 0 .
x1
Since y is arbitrary, we have
f d f
= 0.
y dx y
This is the Euler-Lagranges differential equation to be satisfied by y(x) for the
extremum of the functional between two points.
Generalization of Theorem (1) : Euler-Lagranges equations for several
dependent variables.
Theorem 1a : Derive the Euler-Lagranges equations that are to be satisfied by twice
differential functions y1 , y2 ,..., yn that extremize the integral
x2
I= f ( x, y , y ,..., y , y, y ,..., y )dx

x1
1 2 n 1 2 n
with respect to those functions y1 , y2 ,..., yn which achieve prescribed values at the
fixed points x1 , x2 .
Proof: The functional which is to be extremized can be written as
x2
I= f ( x, y , y )dx,
x1
i i i = 1, 2,..., n .
Choose the family of neighboring curves as

yi ( x, ) = yi ( x, 0 ) + i ( x )

and repeating the procedure delineated either in the Theorem (1) or in the alternate
proof we arrive the following set of Euler-Lagranges equations
f d f
= 0, i = 1, 2,..., n .
yi dx yi
Geodesic : Geodesic is defined as the curve of stationary (extremum) length between

two points.
Worked Examples
Example 1 : Show that the geodesic (shortest distance between two points) in a
Euclidian plane is a straight line.
Solution: Take P ( x1 , y1 ) and Q ( x2 , y2 ) be two fixed points in a Euclidean plane.
Let y = f ( x ) be the curve between P and Q. Then the element of distance between
two neighboring points on the curve y = f ( x ) joining P and Q is given by
ds 2 = dx 2 + dy 2
Hence the total distance between the point P and Q along the curve is given by
Q
I = ds
P
x2 1
dy
(1 + y ) dx,
2 2
I= y = . . . (1)
x1
dx
Here the functional I is extremum if the integrand

1
f = (1 + y2 ) 2 . . . (2)
must satisfy the Euler-Lagranges differential equation

f d f
=0 . . . (3)
y dx y
Now from equation (2) we find that

f f y
= 0 and =
y y 1 + y 2
d y
= 0.
dx 1 + y2

Integrating we get
y = c 1 + y 2 .
Squaring we get
c
y = c1 , where c1 = .
1+ c2
Integrating we get
y = c1 x + c2 . . . . (4)
This is the required straight line. Thus the shortest distance between two points in a
Euclidean plane is a straight line.
Example 2 : Show that the shortest distance between two polar points in a plane is a
straight line.
Solution: Define a curve in a plane. If A ( x, y ) and B ( x + dx, y + dy ) are
infinitesimal points on the curve, then an element of distance between A and B is

given by
ds 2 = dx 2 + dy 2 . . . . (1)
Let = ( r ) be the polar equation of the curve and P ( r1 , 1 ) and Q ( r2 , 2 ) be two
polar points on it. Recall the relations

x = r cos ,
y = r sin .
ds 2 = dr 2 + r 2 d 2 . . . . (2)
Thus the total distance between the points P and Q becomes

r2 1
d
I = (1 + r 2 2 ) 2 dr , = . . . . (3)
r1
dr
The functional I is shortest if the integrand

1
f = (1 + r 2 2 ) 2 . . . (4)
must satisfy the Euler-Lagranges differential equation

f d f
= 0, (5)
dr
d r 2
=0,
dr 1 + r 2 2
r 2 = h 1 + r 2 2 .
Squaring and solving for we get
d h
= 1
.
dr
r (r h
2
)
2 2
On integrating we get
h
= cos 1 + 0 ,
r
where 0 is a constant of integration. We write this as
h = r cos( 0 ) . . . . (6)
This is the polar form of the equation of straight line. Hence the shortest distance
between two polar points is a straight line.
Note : If r = r ( ) is the polar equation of the curve, then the length of the curve is
given by
1 2
dr
I=
0
r2 + d .
d
Since the integrand f = r 2 + r 2 does not contain , we therefore have

f
f r =h.
r
Solving this equation we readily obtain the same polar equation of straight line as the
geodesic.
Example 3 : Show that the geodesic = ( ) on the surface of a sphere is an arc of
the great circle.

Solution : Consider a sphere of radius r described by the equations
x = r sin cos ,
y = r sin sin , . . . (1)
z = r cos .
If A ( x, y, z ) and B ( x + dx, y + dy, z + dz ) be two neighboring points on the curve
joining the points P and Q. Then the infinitesimal distance between A and B along
the curve is given by
ds 2 = dx 2 + dy 2 + dz 2 , . . . (2)
where from equation (1) we find
dx = r cos cos d r sin sin d ,
dy = r cos sin d + r sin cos d , . . . (3)
dz = r sin d .
Squaring and adding these equations we readily obtain
ds 2 = r 2 d 2 + r 2 sin 2 d 2 . . . . (4)
Hence the total distance between the points P and Q along the curve = ( ) is
given by
2 1
d
I = r (1 + sin 2 2 ) 2 d , = . . . (5)
1 d
where
1
f = r (1 + sin 2 2 ) 2 . . . . (6)

The curve is geodesic if the functional I is stationary. This is true if the function f
must satisfy the Euler-Lagranges equations.
f d f
=0 . . . (7)
d
d r sin 2
= 0.
d 1 + sin 2 2

Integrating we get
sin 2 2
= c,
1 + sin 2 2
Solving for we get
c cos ec 2
= 1
.
(1 c 2
cos ec )
2 2
On simplifying we get
d k cos ec 2
= 1
. . . . (8)
d
(1 k 2
cot ) 2 2
Put k cot = t cos ec 2 d = dt ,

Therefore we have
dt
d = .
1 t2
Integrating we get
= sin 1 t ,
or = sin 1 ( k cot ) ,
k cot = sin ( ) ,
k cos = sin sin cos cos sin sin ,
kz = x sin y cos . . . . (9)

This is the first-degree equation in x, y, z, which represents a plane. This plane
passes through the origin, hence cutting the sphere in a great circle. Hence the
geodesic on the surface of a sphere is an arc of a great circle.
Example 4 : Show that the curve is a catenary for which the area of surface of
revolution is minimum when revolved about y-axis.
Solution: Consider a curve between two points ( x1 , y1 ) and ( x2 , y2 ) in the xy plane
y
whose equation is y = y ( x ) . We form a surface
by revolving the curve about y-axis. Our claim is

x
A ds to find the nature of the curve for which the
surface area is minimum. Consider a small strip
at a point A formed by revolving the arc length
x
O ds about y axis. If the distance of the point A
on the curve from y-axis is x, then the surface
z
area of the strip is equal to 2 x ds .
But we know the element of arc ds is given by
ds = 1 + y2 dx .
Thus the surface area of the strip ds is equal to
2 x 1 + y2 dx .
Hence the total area of the surface of revolution of the curve y = y ( x ) about y- axis
is given by
x2
I = 2 x 1 + y2 dx . . . . (1)
x1
This surface area will be minimum if the integrand
f = 2 x 1 + y2 . . . (2)
must satisfy Euler-Lagranges equation

f d f
= 0, . . . (3)
y dx y
d 2 x y
= 0,
dx 1 + y2

d x y
= 0.
dx 1 + y2

Integrating we get
xy = a 1 + y2 .
Solving for y we get

dy a
= .
dx x2 a2
Integrating we get
x
y = a cosh 1 + b .
a
y b
Or x = a cosh . . . . (4)
a
This shows that the curve is the catenary.
The Brachistochrone Problem :
The Brachistochrone is the curve joining two points not lie on the vertical
line, such that the particle falling from rest under the influence of gravity from higher
point to the lower point in minimum time. The curve is called the cycloid.
Example 5: Find the curve of quickest decent.
Or
A particle slides down a curve in the vertical plane under gravity. Find the curve
such that it reaches the lowest point in shortest time.

Solution: Let A and B be two points on the curve not lie on the vertical line. Let
ds
v= be the speed of the particle along the curve. Then the time required to fall an
dt
arc length ds is given by
ds
A
dt =
y v
v 1 + y 2
dt = dx.
v
B Therefore the total time required for the particle to go
x from A to B is given by
B
1 + y 2
t AB = dx . . . (1)
A
v
Since the particle falls freely under gravity, therefore its potential energy goes on
decreasing and is given by
V = mgx ,
and the kinetic energy is given by
1 2
T= mv .
2
Now from the principle of conservation of energy we have
T + V = constant.
Initially at point A, we have x = 0 and v = 0 . Hence the constant is zero.
1 2
mv = mgx ,
2
v = 2 gx . . . . (2)
x2
1 + y 2
t AB =
x1 2 gx
dx . . . . (3)
Thus t AB is minimum if the integrand

1 + y 2
f = , . . . (4)
2 gx
must satisfy Euler-Lagranges equation
f d f
=0 . . . (5)
y dx y
d y
=0
dx 2 gx(1 + y2 )

d y
=0
dx x(1 + y2 )

Integrating we get
y = c x(1 + y2 ) .
Solving it for y we get
dy x
=
dx ax
Integrating we get
x
y= dx + b . . . (6)
ax
Put
x = a sin 2 ( / 2)
. . . . (7)
dx = 2a sin( / 2) cos( / 2)d
Hence
y = a sin 2 ( / 2)d + b .
a
y= ( sin ) + b ,
2
If y = 0, = 0 b = 0 ,
hence

a
y= ( sin ) . . . . (8)
2
Thus from equations (7) and (8) we have
x = b (1 cos ) ,
a
y = b ( sin ) , for b =
2
This is a cycloid. Thus the curve is a cycloid for which the time of decent is
minimum.
Example 6 : Find the extremal of the functional

2
( y y 2 + 2 xy )dx
2
subject to the conditions that

y ( 0 ) = 0, y = 0 .
2
Solution: Let the functional be denoted by

2
I = ( y2 y 2 + 2 xy )dx . . . . (1)
0
The functional is extremum if the integrand

f = y2 y 2 + 2 xy . . . (2)
must satisfy the Euler-Lagranges equation
f d f
= 0, . . . (3)
y dx y
d
2( x y) ( 2 y ) = 0 ,
dx
y + y = x . . . . (4)

This is second order differential equation, whose complementary function (C.F.) is
given by
y = c1 cos x + c2 sin x . . . (5)
where c1 and c2 are arbitrary constants.

The particular integral (P.I.) is
1
y = (1 + D 2 ) x
y = x.
Hence the general solution is given by
y = c1 cos x + c2 sin x + x . . . . (6)
This shows that the extremals of the functional are the two-parameter family of
curves. On using the boundary conditions we obtain
y ( 0 ) = 0 c1 = 0,

y = 0 c2 = .
2 2
Hence the required extremal is

y = x sin x. . . . (7)
2
Example 7 : Find the extremal of the functional

2
x3
1 y2 dx
y (1) = 0, y ( 2) = 3 .
Solution: Let the functional be denoted by

x3
2
I = 2 dx . . . (1)
1
y
The functional is extremum if the integrand

x3
f = . . . (2)
y 2
f d f
= 0. (3)
y dx y
d x3
= 0.
dx y3
Integrating we get
x3 = cy3
or y = ax .
Integrating we get
a 2
y= x +b. . . . (4)
2
Now using the boundary conditions we get
a
y (1) = 0 + b = 0,
2
y ( 2 ) = 3 2a + b = 3.
Solving these two equations we obtain

a = 2, b = 1 .
Hence the required functional becomes
y = x2 1 . . . . (5)
Example 8 : Show that the time taken by a particle moving along a curve y = y ( x )
ds
with velocity = x, from the point (0,0) to the point (1,1) is minimum if the curve
dt
is a circle having its center on the x-axis.
Solution: Let a particle be moving along a curve y = y(x) from the point (0, 0) to the
point (1, 1) with velocity

ds
x=
dt
ds
dt = .
x
Therefore the total time required for the particle to move from the point (0, 0) to the
point (1, 1) is given by
1
ds
t= . . . (1)
0
x
where the infinitesimal distance between two points on the path is given by
dy
ds = 1 + y2 dx, y = .
dx
Hence the equation (1) becomes
1
1 + y 2
t= dx . . . (2)
0
x
Time t is minimum if the integrand
1 + y 2
f = . . . (3)
x
f d f
= 0. . . . (4)
y dx y
d y
= 0,
dx x 1 + y2

y = cx 1 + y2 .
x 1
y = , for a= .
a 2 x2 c
Integrating we get

x
y= dx + b
a x2
2
Put x = a sin dx = a cos d . . . (5)

Therefore,
y = a sin d + b ,
y = a cos + b . . . . (6)
Squaring and adding equations (5) and (6) we get
2
x2 + ( y b ) = a 2 ,
which is the circle having center on y axis.
Example 9 : Show that the geodesic on a right circular cylinder is a helix.

Solution: We know the right circular cylinder is characterized by the equations
x2 + y2 = a 2 , z = z . . . (1)
The parametric equations of the right circular cylinder are
x = a cos ,
y = a sin ,
z = z.
where a is a constant. The element of the distance (metric)
ds 2 = dx 2 + dy 2 + dz 2
on the surface of the cylinder becomes
ds 2 = a 2 d 2 + dz 2 ,
Hence the total length of the curve on the surface of the cylinder is given by

dz
s=

a 2 + z 2 d , for z =
d
. . . . (2)
0
For s to be extremum, the integrand
f = a 2 + z 2 . . . . (3)

f d f
= 0. . . . (4)
z d z
d z
=0.
d f
Integrating the equation and solving for z we get
z = a (constant).
Integrating we get
z = a + b, a 0 , . . . (5)
where a, b are constants. Equation (5) gives the required equation of helix. Thus the
geodesic on the surface of a cylinder is a helix.
Example 10 : Find the differential equation of the geodesic on the surface of an

inverted cone with semi-vertical angle .
Solution: The surface of the cone is characterized by the equation
x 2 + y 2 = z 2 tan 2 , = const. . . . (1)
The parametric equations of the cone are given by
x = ar cos ,
y = ar sin , . . . (2)
z = br.
where for a = sin , b = cos are constant.
Thus the metric ds 2 = dx 2 + dy 2 + dz 2

on the surface of the cone becomes
ds 2 = dr 2 + a 2 r 2 d 2 . . . (3)
Hence the total length of the curve = ( r ) on the surface of the cone is given by
d
s = 1 + a 2 r 2 2 dr , = . . . (4)
dr

The length s is stationary if the integrand
f = 1 + a 2 r 2 2 . . . (5)
f d f
= 0. . . . (6)
dr
d a 2 r 2
= 0,
dr f
Solving for we get

d c1
= , . . . (7)
dr ar a 2 r 2 c12
where c1 = constant.
This is the required differential equation of geodesic, and the geodesic on the surface
of the cone is obtained by integrating equation (7). This gives
c1
= dr + .
ar a r c12
2 2
1 ar
= sec 1 + ,
a c1
c1
r= sec a ( ) .
a
Example 11 : Find the curve for which the functional

4
I y ( x ) = ( y 2 y2 )dx
0
can have extrema, given that y(0)=0, while the right hand end point can vary along

the line x = .
4

Solution: To find the extremal curve of the functional

4
I y ( x ) = ( y 2 y2 )dx , . . . (1)
0
we must solve Eulers equation

f d f
= 0, . . . (2)
y dx y
where f = y 2 y2 . . . (3)
y + y = 0 . . . . (4)
This is the second order differential equation, whose solution is given by
y = a cos x + b sin x . . . . (5)
The boundary condition y ( 0 ) = 0 gives a = 0.
y = b sin x . . . . (6)

The second boundary point moves along the line x = .
4
f
=0
y x =
4
( y) = 0,
x=
4

where from equation (6) we have y = b cos x . Thus y at x = gives
4
b= 0. This implies the extremal is attained on the line y = 0.
Example 12 : If f satisfies Euler-Lagranges equation
f d f
= 0.
y dx y
dg
Then show that f is the total derivative of some function of x and y and
dx
conversely.

Solution: Given that f satisfies Euler-Lagranges equation
f d f
= 0. . . . (1)
y dx y
dg
We claim that f = ,
dx
where g = g ( x, y ) .
As f = f ( x, y , y ) ,
we write equation (1) explicitly as

f 2 f 2 f 2 f
y 2 2 y = 0 . . . . (2)
y xy yy y
We see from equation (2) that the first three terms on the l. h. s. of (2) contain at the
highest the first derivative of y. Therefore equation (2) is satisfied identically if the
coefficient of y vanishes identically.
2 f
= 0.
y2
Integrating w. r. t. y we get
f
= q ( x, y ) .
y
Integrating once again we get
f = q ( x, y ) y + p ( x, y ) , . . . (3)
where p ( x, y ) and q ( x, y ) are constants of integration and may be function of x and
y only. Then the function f so determined must satisfy the Euler Lagranges
equation (1). From equation (3) we find
f q p
= y + ,
y y y
f
and = q ( x, y ) .
y
Therefore equation (1) becomes
q p d
y + ( q ( x, y ) ) = 0 .
y y dx
q p q q
y + y = 0 .
y y x y
p q
= . .(4)
y x
This is the condition that the equation pdx + qdy is an exact differential
equation dg .
dg = pdx + qdy,
dg
= p + qy = f by (3)
dx
Therefore,
dg
f = . . . . (5)
dx
This proves the necessary part.
dg
Conversely, assume that f = . We prove that f satisfies the Euler-Lagranges
dx
equation
f d f
= 0.
y dx y
dg g g
Since f = = + y ,
dx x y
Therefore, we find
f 2 g 2 g
= + y,
y xy y 2
g g
= .
y y

Consider now
f d f 2 g 2 g d g
= + 2 y .
y dx y xy y dx y
f d f 2 g 2 g 2 g 2 g
= + y y .
y dx y xy y 2 xy y 2
f d f
=0
y dx y
dg
f =
dx
satisfies Euler-Lagranges equation.
Example 13 : Show that the Euler-Lagranges equation of the functional

x2
I ( y ( x )) = f ( x, y, y)dx
x1
f
has the first integral f y = const , if the integrand does not depend on x.
y
Solution: The Euler-Lagranges equation of the functional
x2
I ( y ( x )) = f ( x, y, y)dx
x1
to be extremum is given by
f d f
= 0. . . . (1)
y dx y
f 2 f 2 f 2 f
y 2 2 y = 0
y xy yy y
f
If f does not involve x explicitly, then = 0.
x
Therefore, we have

f 2 f 2 f
y 2 2 y = 0 . . . . (2)
y yy y
Multiply equation (2) by y we get
f 2 f 2 f
y y2 2 2 yy = 0 . . . (3)
y yy y
But we know that
d f f f f 2
2 f 2 f
f y = y + y y y y y ,
dx y y y y yy y2
d f f 2
2 f 2 f
f y = y y y y . . . . (4)
dx y y yy y2
From equations (3) and (4) we see that
d f
f y = 0 ,
dx y
f
f y = const. . . . (5)
y
This is the first integral of Euler-Lagranges equation, when the functional
f = f ( y, y ) .
Worked Examples
Example 14 : Find the minimum of the functional

1
1
I ( y ( x) ) = y2 + yy + y + y dx
0
2
if the values at the end points are not given.

Solution: For the minimum of the functional
1
1
I ( y ( x) ) = y2 + yy + y + y dx . . . (1)
0
2
the integrand
1 2
f = y + yy + y + y . . . (2)
2
f d f
= 0. . . . (3)
y dx y
d
y + 1 ( y + y + 1) = 0 ,
dx
y = 1 . . . . (4)
Integrating we get
y = x + c1 , . . . (5)
Further integrating we get
x2
y= + c1 x + c2 , . . . (6)
2
where c1 , c2 are constants of integration and are to be determined.
However, note that the values of y at the end points are not prescribed. In this case
the constants are determined from the conditions.
f f
= 0, and = 0. . . . (7)
y x = 0 y x =1
These two conditions will determine the values of the constants.
( y + y + 1) x =0 = 0, and ( y + y + 1) x=1 = 0 , . . . (8)
where from equation (5) and (6) we have

y ( 0 ) = c1 and y ( 0 ) = c2 ,
1
similarly, y (1) = 1 + c1 and y (1) = + c1 + c2 .
2
Thus the equations (8) become
5
c1 + c2 + 1 = 0, 2c1 + c2 + = 0.
2

3 1
Solving these equations for c1 and c2 we obtain c1 = and c2 =
2 2
Hence the required curve for which the functional given in (1) becomes minimum is
1 2
y=
2
( x 3x + 1) . . . . (9)
Theorem 3 : Find the Euler- Lagrange differential equation satisfied by four times
differentiable function y(x) which extremizes the functional
x2
I ( y ( x )) = f ( x, y, y, y)dx
x1
under the conditions that both y and y are prescribed at the end points.
Proof : Let P ( x1 , y1 ) and Q ( x2 , y2 ) be two fixed points in xy plane. The points P
and Q can be joined by infinitely many curves. Accordingly the value of the integral
I will be different for different paths. We shall look for a curve along which the
functional I has an extremum value. Let c be a curve between P and Q whose
equation is given by y = y(x, 0). Let also the value of the functional along the curve c
be extremum and is given by
x2
I ( y ( x )) = f ( x, y, y, y)dx
x1
. . . (1)
We can label all possible paths starting from P and ending at Q by the family of
equations
y ( x, ) = y ( x, 0 ) + ( x ) , . . . (2)
where is a parameter and ( x ) is any differentiable function of x.
For different values of we get different curves. Accordingly the value of

the integral I will be different for different paths. Since y and y are prescribed at
the end points, this implies that there is no variation in y and y at the end points.

i.e., all the curves of the family and their derivative must be identical at fixed points
P and Q.
This implies that
y ( x1 , ) = y ( x1 , 0 ) = y1 ,
y ( x2 , ) = y ( x2 , 0 ) = y2 ,
and also
y ( x1 , ) = y ( x1 , 0 ) = y1,
y ( x2 , ) = y ( x2 , 0 ) = y2
( x1 ) = 0 = ( x2 ) , . . . (3a)
and ( x1 ) = 0 = ( x2 ) . . . . (3b)
Conversely, the conditions (3) ensure us that the curves of the family that all pass
through the points P and Q. Let the value of the functional along the neighboring
curve be given by
x2
I ( y ( x, ) ) = f ( x, y ( x, ) , y ( x, ) , y ( x, ) ) dx .
x1
. . . (4)
I
From differential calculus, we know the integral I is extremum if = 0,
= 0
because for = 0 the neighboring curve coincides with the curve which gives
extremum values of I . Thus
I
= 0,
= 0
x2
f f f

x1
( x ) + ( x ) + ( x )
y y y
dx = 0 .

Integrating the second and the third integrations by parts, we get

x2 x2 x x2
f f 2
d f f
x y ( x ) dx +
y
( x ) ( x ) dx + ( x )
x1 x1 dx y y x1
1
x2
. . . (5)
d f x2
d 2 f
( x ) + 2 ( x )dx = 0,
dx y x1 x1 dx y
x2
x2
f f d f x2
d f
x y ( ) y dx y ( ) x dx y ( x ) dx +
x dx + x
1 1 x1
. . . (6)
x2 x2
f d f 2
+ ( x) + 2 ( x ) dx = 0.
y x1 x1 dx y
As y and y are both prescribed at the end points, hence on using equations (3) we
obtain
x2
f d f d 2 f
y dx y + dx 2 y ( x ) dx = 0 .
x1
. . . (7)

f d f d 2 f
+ = 0. . . . (8)
y dx y dx 2 y
This is required Euler- Lagrange differential equation to be satisfied by y(x) for
which the functional I has extremum value.
Note : If the functions y and y are not prescribed at the end points then we must
have unlike the fixed end point problem, ( x ) and ( x ) need no longer vanish at
the points x1 and x2 . In order that the curve y =y(x) to be a solution of the variable
end point problem, y must be an extremal. i.e., y must be a solution of Eulers
equation (8). Thus for the extremal we have from equation (6)
x =b x =b
f f d f
y = 0, y dx y = 0. . . . (9)
x =a x =a

Thus to solve the variable end point problem, we must first solve Eulers equation (8)
and then use the conditions (9) to determine the values of the arbitrary constants.
The above result can be summarized in the following theorem (3a).
Theorem (3a) : Derive the differential equation satisfied by four times differentiable
function y ( x ) , which extremizes the integral
x1
I= f ( x, y, y, y)dx
x0
under the condition that both y, y are prescribed at both the ends. Show that if
neither y nor y is prescribed at either end points then
f f
= =0
y x = x0 y x = x1
x = x1
f d f
=0
y dx y x = x0
Remark: (General case of Theorem (3)) If the integrand in equation (1) of the
Theorem (3) is of the form f = f ( x, y, y, y,..., y n ) with the boundary conditions
y ( x1 ) = y1 , y ( x1 ) = y1,....., y n 1 ( x1 ) = y1n 1 ,
y ( x2 ) = y2 , y ( x2 ) = y2 ,....., y n 1 ( x2 ) = y2n 1 ,
then the Euler-Lagranges differential equation is

f d f d 2 f n d
n
f
+ 2 + ..... + ( 1) n = 0.
y dx y dx y dx y n
Worked Examples
Example 15 : Find the curve, which extremizes the functional

4
I ( y ( x ) ) = ( y2 y 2 + x 2 )dx
0

under the conditions that
y ( 0 ) = 0, y ( 0 ) = 1,
1
y = y =
4 4 2
Solution : For the extremization of the functional

4
I ( y ( x ) ) = ( y2 y 2 + x 2 )dx . . . (1)
0
the integrand
f = y2 y 2 + x 2 . . . (2)
f d f d 2 f
+ = 0. . . . (3)
y dx y dx 2 y
d2
2y + ( 2 y ) = 0 ,
dx 2
d4y
i.e., y = 0. . . . (4)
dx 4
The solution of equation (4) is given by
y = ae x + be x + c cos x + d sin x . . . (5)
where a, b, c, d are constants of integration and are to be determined.
Thus
y ( 0 ) = 0 a + b + c = 0,

1 1 1 1
y = ae 4 + be 4 + c+ d= ,
4 2 2 2 2
y ( 0 ) = 1 a b + d = 1,

1 1 1 1
y = ae 4 be 4 c+ d=
4 2 2 2 2

Solving these equations we get a = b = c = 0 and d = 1 .
Hence the required curve is y = sin x .
Example 16 : Minimize the functional

2
1

2
I= ( x ) dt ,
20
satisfies
x ( 0 ) = 1, x ( 0 ) = 1, x ( 2 ) = 1, x ( 2 ) = 0 .
Solution : To minimize the functional

2
1
I = (
2
x ) dt , . . . (1)
20
the integral
1 2
f = x
. . . (2)
2
f d f d 2 f
+ = 0. . . . (3)
x dt x dt 2
x
d4x
This implies = 0. . . . (4)
dt 4
Integrating we get
t3 t2
x = c1 + c2 + c3t + c4 . . . . (5)
6 2
where x given in (5) must satisfy the conditions
x ( 0 ) = 1 c4 = 1,
x ( 0 ) = 1 c3 = 1,
x ( 2 ) = 1 4c1 + 6c2 = 3,
x ( 2 ) = 0 2c1 + 2c2 = 1.

Solving for c1 and c2 we get the required functional is
t2
x= + t +1.
4
Example 17 : Find the function on which the functional
1
I ( y ( x ) ) = ( y2 2 xy )dx ,
0
can be extremized such that

1
y ( 0 ) = y ( 0 ) = 0, y (1) =
120
and y (1) is not prescribed.
Solution : For the extremization of the functional

1
I ( y ( x ) ) = ( y2 2 xy )dx . . . (1)
0
the integrand
f = y2 2 xy . . . (2)
f d f d 2 f
+ = 0. . . . (3)
y dx y dx 2 y
d2
2x + ( 2 y ) = 0,
dx 2
d4y
= x.
dx 4
Integrating we obtain
x2
y = +a,
2
x3
y = + ax + b ,
6

4
x a
y = + x 2 + bx + c ,
24 2
x5 a 3 b 2
y= + x + x + cx + d , . . . (4)
120 6 2
where a, b, c, d are constants to be determined. Given that
y ( 0 ) = 0 d = 0,
y ( 0 ) = 0 c = 0,
1
y (1) = a + 3b = 0.
120
Since y (1) is not prescribed. i.e., y at x= 1 is not given, then we have the
condition that
f
= 0 y (1) = 0 .
y x =1
This gives from above equation that
6a + 6b = 1 .
Solving the equations for a and b we get
1 1
a = , and b = .
4 12
Substituting these values in equation (4) we get
x5 1
y= + ( x 2 x3 ) . . . . (5)
120 24
This is the required curve.
When integrand is a function more than two dependent variables :

Example 18: Prove that the shortest distance between two points in a Euclidean 3-
space is a straight line.
Solution: Define the curve y = y ( x ) , z = z ( x ) in the 3-dimentional Euclidean
space. Let P ( x, y, z ) and Q ( x + dx, y + dy, z + dz ) be two neighboring points on the

curve joining the points A ( x1 , y1 , z1 ) and B ( x2 , y2 , z2 ) .Thus the infinitesimal
distance between P and Q along the curve is given by

ds 2 = dx 2 + dy 2 + dz 2 .
Hence the total distance between the points A and B along the curve is given by
x2 1
dy
(1 + y ) dx,
2 2 2
I= + z y = . . . (1)
x1
dx
1
Let f = (1 + y 2 + z 2 ) 2 . . . . (2)
We know the functional I is shortest if the function f must satisfy the Euler-
Lagranges equations.
f d f
= 0, . . . (3)
y dx y
f d f
and = 0. . . . (4)
z dx z
d y
=o y = af , a = constant
dx f
and
y 2 (1 a 2 ) a 2 z 2 = a 2 . . . . (5)
Similarly, from equation (4) we obtain

z 2 (1 b 2 ) b 2 y 2 = b 2 . . . . (6)
Solving equations (5) and (6) for y and z we get

a a
y = , and z = ,
1 a2 b2 1 a 2 b2
1
2 2
i.e., y = c1 , for c1 = a (1 a b 2
) . . . (7)
1
2 2
and z = c2 for c2 = b (1 a b 2
) . . . . (8)

Integrating equations (7) and (8) we get
y = c1 x + ( z ) , . . . (9)
z = c2 x + ( y ) , . . . (10)
where ( z ) and ( y ) are constants of integration and may be functions of z and y
respectively. Thus the required curve is given by equations (9) and (10). But these
equations represent a pair of planes. The common point of intersection of these
planes is the straight line. Hence the shortest distance between two points in
Euclidean 3-space is a straight line.
Example 19 : Show that the geodesic defined in the 3-dimentional Euclidean space
by the equations x = x ( t ) , y = y ( t ) , z = z ( t ) is a straight line.
Solution : Let
x = x (t ) , y = y (t ) , z = z (t ) . . . (1)
be a curve in 3-dimentional Euclidean space, where t is a parameter of the curve. The

infinitesimal distance between two neighboring points on the curve (1) is given by
ds 2 = dx 2 + dy 2 + dz 2 ,
where from equation (1) we have
dx = xdt
, dy = ydt
, dz = zdt
.
Thus
ds 2 = ( x 2 + y 2 + z 2 ) dt 2 .
Hence the total length of the curve between the points P ( t0 ) and P ( t1 ) is given by
1
t1 2
I = ( x 2 + y 2 + z 2 ) dt . . . (2)
t0
The curve is geodesic if the length of the curve I is extremum. This is true if the
integrand

f = x 2 + y 2 + z 2 . . . (3)
must satisfy the Euler-Lagranges equations.
f d f
= 0 i = 1, 2,3 with xi = ( x, y, z ) , . . . (4)
xi dt xi
where
f f x
= 0 and = i i = 1, 2,3 .
xi xi f
x = af , y = bf , z = cf .
Thus the Euler-Lagranges equations become
(a 2
1) x 2 + a 2 y 2 + a 2 z 2 = 0,
b 2 x 2 + ( b 2 1) y 2 + b 2 z 2 = 0, . . . (5)
c 2 x 2 + c 2 y 2 + ( c 2 1) z 2 = 0.
These equations are consistent provided

a2 1 a2 a2
b2 b2 1 b2 = 0
c2 c2 c2 1
a2 + b2 + c2 = 1 .
Solving equations (5) we obtain
a
x = z, . . . (6)
1 a2 b2
b
y = z, z 0 . . . . (7)
1 a 2 b2
Integrating equations (6) and (7) we obtain
a
x = c1 z + ( y ) , c1 = , . . . (8)
1 a 2 b2
b
y = c2 z + ( x ) , c2 = . . . . (9)
1 a2 b2
where ( x ) and ( y ) are constants of integration and may be functions of x and y
respectively. Equations (8) and (9) represent planes. The locus of the common points
of these planes is the straight line. Hence the geodesic in 3-dimentional Euclidean
space is the straight line.
Unit 2: Isoperimetric Problems:

The problems in which the function which is eligible for the extremization of
a given definite integral is required to confirm with certain restrictions that are given
as the boundary conditions. Such problems are called isoperimetric problems. The
method is exactly analogous to the method of finding stationary value of a function
under certain conditions by Lagranges multipliers method.
Theorem 4 : Obtain the differential equation, which is satisfied by the functional
f ( x, y, y ) which extremizes the integral
x2
I ( y ( x)) = f ( x, y, y)dx
x1
subject to the conditions y ( x1 ) = y1 , y ( x2 ) = y2 , and the integral

x2
J = g ( x, y, y )dx = constant.
x1
Proof : Consider the functional between two points P ( x1 , y1 ) and Q ( x2 , y2 ) given
by
x2
I ( y ( x)) = f ( x, y, y)dx
x1
. . . (1)
subject to the conditions

y ( x1 ) = y1 , y ( x2 ) = y2 , . . . (2)

x2
and J = g ( x, y, y )dx = constant. . . . (3)

x1
x) The points P and Q can be joined by infinitely

y
2
2(
+
x) many curves.
1
1(
+ Q(x 2, y 2)
x)
y( Accordingly the value of the integral I will be
Y= y ( x)
y=
different for different paths. Let all possible paths
P(x1 , y 1) starting from P and ending at Q be given by two
x
O parameters family of curves
Y ( x ) = y ( x ) + 11 ( x ) + 22 ( x ) . . . (4)
where 1 , 2 are parameters and 1 ( x ) ,2 ( x ) are arbitrary differentiable functions of
x such that
1 ( x1 ) = 0 = 1 ( x2 ) ,
. . . (5)
2 ( x1 ) = 0 = 2 ( x2 )
These conditions ensure us that the curves of the family that all pass through the
points P and Q.
Note that, we can not however, express Y(x) as merely a one parameter family of
curves, because any change in the value of the single parameter would in general
alter the value of J, whose constancy must be maintained as prescribed. For this
reason we introduce two parameter families of curves. We shall look for a curve
along which the functional I has an extremum value under the condition (3). Let c
be such a curve between P and Q whose equation is given by y = y (x) such that the
functional (1) along the curve c has extremum value. The values of the integrals (1)
and (3) along the neighboring curve (4) are obtained by replacing y by Y in both the
equations (1) and (3). Thus we have
x2
I ( 1 , 2 ) = f ( x, Y , Y ) dx
x1
. . . (6)

x2
and J ( 1 , 2 ) = g ( x, Y , Y ) dx = const. . . . (7)

x1
Equation (7) shows that 1 and 2 is not independent, but they are related by
J ( 1 , 2 ) = const. . . . (8)
Thus the changes in the value of the parameters are such that the constancy of (7) is
maintained. Thus our new problem is to extremizes (6) under the restriction (7). To
solve the problem we use the method of Lagranges multipliers.
Multiply equation (7) by and adding it to equation (6) we get
x2
I ( 1 , 2 ) = I + J = f ( x, Y , Y )dx ,
* *
. . . (9)
x1
where is Lagranges undetermined multiplier and

f * = f + g . . . (10)
Thus extremization of (1) subject to the condition (3) is equivalent to the
extremization of (9). Thus from differential calculus, the integral I * is extremum if
I *
= 0.
j 1 = 0, 2 =0
Thus from equation (9) we have
I * x2
f * f *
=0 j ( x ) +
y
j ( x ) dx = 0 . J=1, 2 . . . (11)
j 1 = 0, 2 =0 x1 y
Note here that by setting 1 = 2 = 0 , we replace Y and Y to y, y .

Integrating the second integration by parts, we get
x2 x2 x
f * f * 2
d f *
x y j ( ) y j ( ) x dx y j ( x ) dx = 0 .
x dx + x . . . (12)
1 x 1 1
As any curve is prescribed at the end points, hence on using conditions (5) we obtain

x2
f * d f *
y dx y j ( x ) dx = 0, j = 1, 2 .
x1
f * d f *
= 0. . . . (13)
y dx y
This is required Euler- Lagranges differential equation to be satisfied by y (x) for

which the functional I has extremum value under the condition (3).
Remark : If y is not prescribed at either end point then from equation (12) we have
f *
= 0 at that end point.
y
Generalization of Theorem 4 : Euler-Lagranges equations for several

dependent variables :
Theorem 4a : Obtain the differential equations which must be satisfied by the
function which extremize the integral
x2
I= f ( x, y , y ,..., y , y, y ,..., y )dx

x1
1 2 n 1 2 n
with respect to the twice differentiable functions y1 , y2 ,..., yn for which

x2
J = g ( x, y1 , y2 ,..., yn , y1, y2 ,..., yn )dx = const.

x1
and with y, yi prescribed at points x1 , x2 .

Proof : The functional which is to be extremized can be written as
x2
I= f ( x, y , y )dx,
x1
i i i = 1, 2,..., n
together with the conditions

x2
J = g ( x, yi , yi )dx = const. i = 1, 2,..., n

x1
and y, yi prescribed at points x1 , x2 .

Repeating the procedure described in the Theorem (4) we arrive the following set of
Euler-Lagranges equations
f * d f *
= 0, i = 1, 2,..., n
yi dx yi
where f * = f + g .
Theorem 5 : Obtain the differential equation, which is satisfied by four times
differential function y (x) which extremizes the functional
x2
I ( y ( x)) = f ( x, y, y, y)dx
x1
subject to the conditions that the integral

x2
J = g ( x, y, y, y )dx = constant.
x1
and both y and y are prescribed at the end points.

Proof: Proof of the Theorem 5 runs exactly in the same manner as that of the proof
of Theorem 3 and Theorem 4. The required Euler-Lagranges differential equation in
this case is given by
f * d f * d 2 f *
+ = 0,
y dx y dx 2 y
where f * = f + g .

Remarks :
1. If y is not prescribed at either end point, then we have the condition
f * d f *
= 0 at that end point.
y dx y
f *
2. If y is not prescribed at either end point then we have = 0 at that point.
y
3. In general if
f = f ( x, y, y, y,..., y n )
g = g ( x, y, y, y,..., y n )
with the boundary conditions that y, y, y,..., y n 1 are prescribed at both the
ends, then in this case the Euler-Lagranges equation is
f * d f * d 2 f * n d
n
f *
+ + ... + ( ) n n =0.
1
y dx y dx 2 y dx y
Worked Examples
Example 20 : Find the plane curve of fixed perimeter that encloses maximum area.
(The problem is supposed to have arisen from the gift of a king who was
happy with a person and promised to give him all the land that he could enclose by
running round in a day. The perimeter of his path was fixed.)
Solution: Let c : y = y (x) be a plane curve of fixed perimeter l .
x2
l = ds , . . . (1)
x1
where the infinitesimal distance between two points on the curve is given by
dy
ds = 1 + y2 dx, y = .
dx

Hence the total length of the curve between two points P and Q becomes
x2

x1
1 + y2 dx = l . . . . (2)
The area bounded by the curve c and the x-axis is given by

x2
A ( y ( x )) = ydx .
x1
. . . (3)
Thus we maximize (3) subject to the condition (2). Hence the required Euler-
Lagranges equation to be satisfied is
f * d f *
= 0, . . . (4)
y dx y
where
f * = f + g,
. . . (5)
f * = y + 1 + y 2 .
Solving equation (4) we get
d y
1 = 0,
dx 1 + y2

y
x =a.
1 + y 2
2 2 y 2
( x a) = ,
(1 + y )
2
xa
or y = . . . . (6)
2
2 ( x a)
Integrating we get

y=
( x a) dx + b . . . (7)
1
( x a )
2 2 2

Put
x a = sin t ,
. . . (8)
dx = cos tdt
y = sin tdt + b,
y = cos t + b . . . (9)
2 2
( x a) + ( y b) = 2 . . . . (10)
This is a circle centered at (a, b) and of radius and is to be determined. To

l
determine , we know that the circumference of the circle is 2 = l = .
2
Example 21 : Find the shape of the plane curve of fixed length l whose end points
lie on the x-axis and area enclosed by it and the x-axis is maximum.
Solution: Let c: y = y(x) be a plane curve of fixed length l whose end points lie on
the x-axis and the curve lies in the upper half
y
plane. The area bounded by the curve c and
the x-axis is given by
y = y(x) x2
A ( y ( x )) = ydx
x1
. . . (1)
x such that the length of the curve is fixed and

O (x1, 0) (x 2, 0)
is given by
x2 x2
J = ds = 1 + y2 dx = l . . . (2)
x1 x1
The area given in equation (1) is maximum under the condition (2), if

f * d f *
= 0, . . . (3)
y dx y
where
f * = f + g,
. . . (4)
f * = y + 1 + y 2
where is Lagranges multiplier. Solving equation (3) we get
d y
1 = 0,
dx 1 + y2

y
x =a
1 + y 2
2 2 y 2
( x a) = ,
(1 + y )
2
xa
or y = . . . . (5)
2
2 ( x a)
Integrating we get
y=
( x a) dx . . . . (6)
1
2 ( x a ) 2 2

Put
x a = sin t ,
. . . (7)
dx = cos tdt
y = sin tdt + b,
y = cos t + b . . . . (8)
2 2
( x a) + ( y b) = 2 . . . . (9)

Thus the curve is a semi circle centered at (a, b) and of radius , and is to be
determined.
l
Since from the condition (2) the perimeter of the semi-circle is = l = .

This is the radius of the curve of fixed length l which encloses maximum area.
Example 22 : Find the extremals for an isoperimetric problem

1
I ( y ( x ) ) = ( y2 + x 2 ) dx
0

1
y dx = 2, y ( 0 ) = 0, y (1) = 0 .
2
Solution : The functional, which is to be extremized, is given by

1
I ( y ( x ) ) = ( y2 + x 2 ) dx . . . (1)
0
such that
1
y dx = 2
2
. . . (2)
0
and
y ( 0 ) = 0, y (1) = 0 . . . (3)
Thus for the extremizes of (1) under (2), we know the condition to be satisfied is
f * d f *
= 0, . . . (4)
y dx y
where
f * = f + g,
. . . (5)
f * = y 2 + x 2 + y 2 ,
where is Lagranges multiplier.

Solving equation (4) we get
y y = 0 . . . . (6)
This equation has roots for > 0 or i for < 0 .

Case 1 : Let > 0 .
The solution of equation (6) is
x
y = ae x
+ be , . . . (7)
Conditions (3) give
y ( 0 ) = 0 a + b = 0,
y (1) = 0 ae
+ be
= 0.
Solving for a and b we have for

b 0, e 2 =1
e2 + 2 in
=1
2 + 2in = 0,
= in .
This is contradictory to is positive, hence b = 0 and consequently a = 0 proving
that the equation has only trivial solution.
Case 2 : Let < 0 .
The solution of equation (6) is
y = a cos x + b sin x . . . . (8)
Boundary conditions (3) give
y ( 0) = 0 a = 0
y (1) = 0 b sin x = 0,
= n , for 0.
Hence the required solution becomes we get
y = b sin n x . . . . (9)

Condition (2) gives
b = 2 .
Therefore the required solution is y = 2sin n x .
Example 23 : Find the extremals for an isoperimetric problem

I ( y ( x ) ) = ( y2 y 2 ) dx
0

ydx = 1,
0
y ( 0 ) = 0, y ( ) = 1 .
Solution : It is given that

I ( y ( x ) ) = ( y2 y 2 ) dx . . . (1)
0
where the functional I is to be extremized under the conditions

ydx = 1,
0
. . . (2)
and y ( 0 ) = 0, y ( ) = 1 . . . (3)
The corresponding Euler-Lagranges equation is
f * d f *
= 0, . . . (4)
y dx y
where
f * = f + g,
. . . (5)
f * = y 2 y 2 + y
Hence the equation (4) becomes

y + y = . . . . (6)
2
The C.F. of equation (6) is given by
y = a cos x + b sin x ,

where as the P.I. is given by

y= .
2
Hence the general solution becomes

y = a cos x + b sin x + . . . (7)
2
To determine the arbitrary constants of integration we use the boundary
conditions (3)

y ( 0) = 0 a = ,
2
1
y ( ) = 1 = 1, a =
2
To determine other constant of integration we use

ydx = 1,
0

1 1
cos xdx +b sin xdx + dx = 1.
20 0
20
This gives the value of b as

2
b= .
4
Hence the required curve is
1 1
y= (1 cos x ) + ( 2 ) sin x . . . . (8)
2 4
Example 24 : Prove that the extremal of the isoperimetric problem

4
I = y2 dx, y (1) = 3, y ( 4 ) = 24 ,
1
subject to the condition

4
ydx = 36
1
is a parabola.
Solution: Here f * = y2 + y .
The corresponding Euler-Lagranges differential equation is
2 y = 0 . . . . (1)
Integrating two times we get

y= x 2 + ax + b , . . . (2)
4
where the constants of integration are to be determined. Now the boundary
conditions
y (1) = 3 + 4a + 4b = 12,
. . . (3)
y ( 4 ) = 24 4 + 4a + b = 24.
Also the condition

4
ydx = 36 ,
1
gives
4

4 x
2
+ ax + b dx = 36 ,
1
21 + 30a + 12b = 144 . . . . (4)
Solving equations (3) and (4) we obtain
a = 2, b = 0, = 4.
Thus the required curve is obtain by putting these values in equation (2) and is
y = x2 + 2 x .
2
We write this as ( x + 1) = y +1

Or equivalently for X=(x+1), Y=(y+1), we have
X2 =Y .
Hence the curve is a parabola.
Example 25 : Find the extremals for the isoperimetric problem

t
12
I = ( xy yx )dt . . . (1)
2 t1
with the conditions that

t2
J = x 2 + y 2 dt = l , . . . (2)
t1
x ( t1 ) = x ( t2 ) = x0 ,
. . . (3)
y ( t1 ) = y ( t2 ) = y0 .
Solution : We wand to find the function for which equation (1) is extremum w. r. t.
the functions x(t), y(t) satisfying the conditions (2) and (3). We know the conditions
that the integral (1) is extremum under (2) if the following Euler Lagranges
equations are satisfied.
f * d f *
= 0, . . . (4)
x dt x
f * d f *
= 0, . . . (5)
y dt y
where
f * = f + g,
1
f* = ( xy yx ) + x 2 + y 2 . . . . (6)
2
Hence the equations (4) and (5) reduce to
y d y x
+ =0
2 dt 2 x 2 + y 2

x d x y
+ = 0.
2 dt 2 x 2 + y 2

Integrating these equations w. r. t. t we get
y y x
+ =a
2 2 x 2 + y 2

x
ya =
x + y 2
2
x x y
+ + = b,
2 2 x
2
+ y
2
and
y
xb = .
x 2 + y 2
Squaring and adding above equations we get

2 2
( x b) + ( y a) = 2 . . . . (7)
This is a circle of radius and centered at (b, a). Thus the closed curve for which
the enclosed area is maximum is a circle. The length of the circle is
l
2 = l = . This gives the radius of the curve.
2
Exercise:
1. Show that the shortest distance between two points along the curve
x = x (t ) , y = y ( t ) in a Euclidean plane is a straight line.
2. Show that the geodesic defined by r = r ( t ) , = ( t ) in a plane is a straight
line.
3. Show that the stationary (extremum) distance between two points along the
curve = (t ) , = (t ) on the sphere
x = r sin cos , y = r sin sin , z = r cos is an arc of the great circle.

4. Find the geodesic on the surface obtained by generating the parabola
y 2 = 4ax about x-axis.
Ans.: The surface of revolution is y 2 + z 2 = 4ax , whose parametric

representation is
x = au 2 , y = 2au sin v, z = 2au cos v
The geodesic on this surface is obtain by solving the integral
1+ u2
v = c1 du + c2 .
u u 2 c12
5. Derive the Euler-Lagranges equations that are to be satisfied by twice
differential functions x ( t ) , y ( t ) ,..., z ( t ) , that extremize the integral
t2
dx
I = f ( x, y,..., z , x, y ,..., z, t )dt , x =
t1
dt
which achieve prescribed values at the fixed points t1 , t2 .
f d f f d f f d f
Ans : = 0, = 0,..., = 0,
x dt x y dt y z dt z
6. Find the curve which generates a surface of revolution of minimum area
when it is revolved about x axis.
Ans : Area of revolution of a curve about x-axis is
x1
I= y
x0
1 + y2 dx .
xb
The curve is a catenary given by y = c sec , or y = a cosh .
a
7. Find the function on which the functional can be extremized
1
1
I y ( x ) = ( y2 2 xy )dx, y ( 0 ) = 0, y (1) = ,
0
120
and y is not prescribed at both the ends.

x5 x 3 x
Ans: y = + .
120 36 36
4
( xy y )dx , which is determined by

2
8. Find the stationary function of the
0
boundary conditions y ( 0 ) = 0, y ( 4 ) = 3.
x2 x
Ans : y = .
4 4
2
x3
9. Find the extremum of I y ( x ) = 2 dx, y (1) = 1, y ( 2 ) = 4.
1
y
Ans: y = x 2 .
x1
y2
10. Find the extremal of I y ( x ) =
x0
x3
dx.
x4
Ans: y = c1 + c2 .
4

( y y 2 + 4 y cos x )dx, y ( 0 ) = 0, y ( ) = 0.
2
11. Find the extremal of
0
Ans: y = ( c2 + x sin x ) .
12. Find the extremal of the functional

1
( y 12 xy )dx, y ( 0 ) = 1, y (1) = 2.
2
Ans: y = x 3 + 2 x + 1 .
13. Determine the curve z = z ( x ) for which the functional

2
I = ( z 2 2 xz )dx, z (1) = 0, z ( 2 ) = 1 is extremal.
1
x3 x
Ans: z = + .
3 6

14. Determine the curve z = z ( x ) for which the functional

2
I = ( z 2 z 2 )dx, z ( 0 ) = 0, z = 1
0 2
is extremal.
Ans: z = sin x .

2
I y ( x ) = ( y2 y 2 + x 2 )dx, y ( 0 ) = 1, y = 0, y ( 0 ) = 0, y = 1.
0 2 2
Ans : y = cos x. .
16. Obtain the differential equation in which the extremizing function makes the
integral
x2
I ( y ( x)) = f ( x, y, y)dx
x1
extremum subject to the conditions y ( x1 ) = y1 , y ( x2 ) = y2 , and

x2
J k = g k ( x, y, y )dx = const. k = 1, 2,..., n

x1
Ans : The problem would be extrmization of

x2
f ( x, y, y)dx ,
* *
I =
x1
n n
where I * = I + k J k and f * = f + k g k .
k =1 k =1
f * d f *
The required differential equation is = 0.
y dx y

17. Find the extremals of the isoperimetric problem
x1 x1
I=
x0
y2 dx, s.t. ydx = c.
x0
x2
Ans: y = + ax + b .
4
1
I y ( x ) = (1 + y2 )dx, y ( 0 ) = 0, y (1) = 1, y ( 0 ) = 1, y (1) .
0
Ans: y = x .
x1
I y ( x ) , z ( x ) = ( 2 yz 2 y + y2 z 2 )dx.
2
x0
Ans: y = ( c1 x + c2 ) cos x + ( c3 x + c4 ) sin x, z = 2 y + y .

x1
I y ( x ) = (y + y2 + 2 ye x )dx.
2
x0
xe x
Ans: y = + c1e x + c2 e x .
2

CHAPTER - III
HAMILTONS PRINCIPLE AND HAMILTONS

FORMULATION
Unit 1: Hamiltons Principles:

Introduction :
In the Chapter II we have used the techniques of variational principles of
Calculus of Variation to find the stationary path between two points. Hamiltons
principle is one of the variational principles in mechanics. All the laws of mechanics
can be derived by using the Hamiltons principle. Hence it is one of the most
fundamental and important principles of mechanics and mathematical physics.
In this unit we define Hamiltons principle for conservative and non-
conservative systems and derive Hamiltons canonical equations of motion. We also
derive Lagranges equations of motion.
Hamiltons Principle (for non-conservative system) :
Hamiltons principle for non-conservative systems states that The motion of
a dynamical system between two points at time intervals t0 to t1 is such that the line
integral
t1
I = (T + W )dt
t0
is extremum for the actual path followed by the system , where T is the kinetic
energy and W is the work done by the particle.
It is equivalent to say that variation in the actual path followed by the
system is zero. Mathematically, it means that

t1
I = (T + W )dt = 0
t0
for actual path.

Hamiltons Principle (for conservative system) :
Of all possible paths between two points along which a dynamical system
may move from one point to another within a given time interval from t0 to t1 , the
actual path followed by the system is the one which minimizes the line integral of
Lagrangian.
This means that the motion of a dynamical system from t0 to t1 is such that
t1
the line integral Ldt

t0
is extremum for actual path. This implies that small
variation in the actual path followed by the system is zero.

t1
Mathematically, we express this as Ldt = 0 for the actual path.

t0
Note : We will show bellow in the Theorem (2) that the Hamiltons principle
t1
Ldt = 0 also holds good for the non-conservative system.

t0
Action in Mechanics :
Let L = L ( q j , q j , t ) be the Lagrangian for the conservative system. Then the
integral
t1
I = Ldt
t0
is called the action of the system.

Hence we can also define the Hamiltons principle as Out of all possible
paths of a dynamical system between the time instants t0 and t1 , the actual path
followed by the system is one for which the action has a stationary value

t1
I = Ldt = 0
t0
for the actual path.

Theorem 1 : Derive Hamiltons principle for non-conservative system from
DAlemberts principle and hence deduce from it the Hamiltons principle for
conservative system.
Proof: We start with DAlemberts principle which states that
( F p ) r = 0 .
i
i i i . . . (1)
Note that in this principle the knowledge of force whether it is conservative or non-
conservative and also the requirement of holonomic or non-holonomic constraints
does not arise. We write the principle in the form
F r = p r .
i
i i
i
i i
W = p i ri . . . . (2)
i
where W = Fi ri is the virtual work.

i
Now consider
p r = m r r ,
i
i i
i
i i i
d d
= ( mi ri ri ) mi ri ( ri ) .
i dt i dt
d
Since we have ri = ri , therefore, we write
dt
d 1
p r = dt m r r 2 m r
i
i i
i
i i i
i
i i
2
.

d
p r = dt m r r T ,
i
i i
i
i i i

1
where T = mi ri 2
i 2
is the kinetic energy of the system. Substituting this in equation (2) we get
d
W =
dt i
mi ri ri T

d
(W + T ) =
dt i
mi ri ri .

Integrating the above equation with respect to t between t0 to t1 we get
t1 t1

(W + T )dt = mi ri ri .
t0 i t0
Since, there is no variation in co-ordinates along any paths at the end points. i.e.,
t1
( ri )t
0
= 0 . Hence from above equation we have
t1
(W + T )dt = 0 . . . . (3)
t0
This is known as Hamiltons principle for non-conservative systems.

If however, the system is conservative, then the forces are derivable from potential.
In this case the expression for virtual work becomes
V
W = Fi ri = ri = V .
i i ri
t1
(T V )dt = 0,
t0
t1
Ldt = 0. . . . (4)
t0
This is the required Hamiltons principle for conservative system.

t1
Theorem 2 : Show that the Hamiltons principle Ldt = 0 also holds for the
t0
non- conservative system.

Proof : We know the Hamiltons principle for non-conservative system is given by
t1
(T + W )dt = 0 . . . (1)
t0
for actual path. The expression for the virtual work is given by
ri
W = Fi ri = Fi q j
i i j q j
.
r
W = Fi i q j
i
j q j
W = Q j q j , . . . (2)
j
ri
where Q j = Fi
i q j
are the components of generalized forces. In the case of non-conservative system the
potential energy is dependent on velocity called the velocity dependent potential. In
this case the generalized force is given by
U d U
Qj = + .
q j dt q j
Substituting this in equation (2) and integrating it between the limits t0 to t1 we find
t1 t1 t1 U d U
Wdt = Q j q j dt = q + q j dt ,
dt q j
t0 t0 j t0 j j
t1
t1 t1
U d U
Tdt = q j
q j dt j dt q q j dt .
t0 t0 j
t0 j

Integrating the second integral by parts we obtain
t
t1 t1
U U
1 t1
U d
t Tdt =
q j q j dt

j q
q j
+ q ( q j ) dt .
0 t0 j j t0 t0 j j dt
t1
Since change in co-ordinates at the end point is zero. ( q j )t = 0
0
d
and also
dt
( q j ) = q j ,
then we have
t1 t1
U U
Tdt = j q j q j dt .
q
+ q
t0 t0
j j
Since time t is fixed along any path hence, there is no variation in time along any
path therefore change in time along any path is zero. i.e., t = 0
Hence we write above equation as
t1 t1
U U U
Tdt = j q j q j + t t dt .
q
+ q
t0 t0
j j
t1 t1
Tdt = Udt .
t0 t0
t1
(T U ) dt = 0,
t0
t1
Ldt = 0.
t0
This proves that the Hamilton's principle holds good even for non-conservative
systems.
Theorem 3 : State Hamiltons principle for non-conservative system and hence

derive Lagranges equations of motion for non-conservative holonomic systems.

Proof: Let us consider a non-conservative holonomic dynamical system whose
configuration at any instant t is specified by n generalized co-ordinates
q1 , q2 , q3 ,..., qn . Hamiltons principle for non-conservative system states that
t1
(T + W )dt = 0 for actual path. . . . (1)

t0
The virtual work done is given by

ri
W = Fi ri = Fi q j
i i j q j
.
r
W = Fi i q j
j
i q j
W = Q j q j , . . . (2)
j
ri
where Q j = Fi
i q j
are the components of generalized forces.

t1 t1
Wdt = Q j q j dt .
t0 t0 j
. . . (3)
The kinetic energy of the particle T = T ( q j , q j , t ) .
T T T
T = q j + q j + t . . . . (4)
j q j j q
j t
As the variation in time along any path is always zero. t = 0 .

This implies that
T T
T = q j + q j . . . (5)
j q j j q
j
Integrating equation (5) between the limits t0 to t1 we get

t1 t1 t
T 1
T
Tdt =
t0 t0 j q j
q j dt +
t0 j q j
q j dt

Since we have
d
q j = q j .
dt
Therefore we write the above integral as
t1 t1 t
T 1
T d
Tdt =
t0 t0 j q j
q j dt +
t0 j
q j dt
( q j ) dt .
Integrating the second integral by parts, we obtain
t1
t1 t1
T T t1
d T
Tdt = q dt + q
q j j j q j j t j dt q j
q j dt .
t0 t0 j 0 t0
Since in variation there is no change in the co-ordinates at the end points

t1
( q )
j t
0
= 0 . Hence
t1 t1
T d T
Tdt = q j

dt q j
q j dt . . . . (6)
t0 t0 j
t1
T d T
q
dt q j
+ Q j q j dt = 0 . . . . (7)
t0 j
j
If the constraints are holonomic then q j are independent. (Note that if the
constraints are non-holonomic, then q j are not all independent. In this case
vanishing of the integral (7) does not imply the coefficient vanish separately) Hence
the integral (7) vanishes if and only if the coefficient must vanish separately.
d T T
= Qj . . . . (8)
dt q j q j
These are the Lagranges equations of motion for non-conservative holonomic
system.

Theorem 4 : Deduce Hamiltons principle for conservative system from
DAlemberts principle.
Proof: We start with DAlemberts principle which states that
( F p ) r = 0 .
i
i i i . . . (1)
We write the principle in the form
F r = p r ,
i
i i
i
i i . . . (2)
where ri is the virtual displacement and occurs at a particular instant. Hence

change in time t along any path is zero.
Now consider
p r = m r r ,
i
i i
i
i i i
d d .
= ( i i i) i i ( i)
m r r m r
r
i dt i dt
d
Since we have ri = ri ,
dt
therefore we write
d 1
p r = dt m r r 2 m r
i
i i
i
i i i
i
i i
2
.

d
p r = dt m r r T
i
i
i
i i i . . . (3)
where
1
T = mi ri 2
i 2
is the kinetic energy of the system. Substituting equation (3) in equation (2) we get
d
F r = dt m r r T
i
i i
i
i i i . . . (4)

V
Since the force is conservative Fi = .
ri
d V

dt i
mi ri ri = T
i ri
ri
.
= T V
Integrating the above equation with respect to t between t0 to t1 we get

t1 t
1
i i i
m r
r = Ldt .
i t0 t0
Since, there is no variation in co-ordinates along any paths at the end points.
t
i.e. ( ri )t1 = 0 . Hence from above equation we have
0
t1
Ldt = 0 . . . . (5)
t0
This is the required Hamiltons principle for conservative systems.

Derivation of Lagranges equations of motion from Hamiltons Principle :
Theorem 5 : Show that the Lagranges equations are necessary conditions for the
action to have a stationary value.
Proof: We know the action of a particle is defined by
t1
I = Ldt . . . (1)
t0
where L is the Lagrangian of the system. Consider

t1
I = Ldt ,
t0
t1
L L
= qj + q j dt
j q j
t0 j q
j
As there is no variation in time along any path, hence t = 0 .

t1 t1 t
L 1
L
Ldt = q j dt + q j dt.
t0 t0 j q j t0 j q j
dq j d
Since
dt
=
dt
( q j ) ,
therefore, we write
t1 t1 t
L 1
L d
Ldt = q j dt + ( q j )dt . . . (2)
t0 t0 j
q j t0 j
q j dt
Integrating the second integral on the r. h. s. of equation (2) we get

t1
t1 t1
L L t1
d L
Ldt = q j dt + qj q j dt.
q j
t0 t0 j j q j t0 j dt q j t0
Since there is no variation in the co-ordinates along any path at the end points, hence
change in the co-ordinates at the end points is zero. i.e.,
t1
( q ) j t
0
= 0.
Thus we have
t1 t1
L d L
Ldt = q j dt. . . . (3)
t0 t0 j q j dt q j
If the system is holonomic, then all the generalized co-ordinates are linearly
independent and hence we have
t1 t1
L d L
Ldt = 0 q j dt = 0
t0 t0 j
q j dt q j
L d L
t1
Ldt = 0 = 0. . . . (4)
t0
q j dt q j
These are the required Lagranges equations of motion derived from the Hamiltons
principle. The equation (4) also shows that the Lagranges equations of motion for
holonomic system are necessary and sufficient conditions for action to have a
stationary value.
Worked Examples
Example 1 : Use Hamiltons principle to find the equations of motion of a particle of

unit mass moving on a plane in a conservative force field.
Solution: Let the force F be conservative and under the action of which the particle
of unit mass be moving on the xy plane. Let P (x, y) be the position of the particle.
We write the force
F = iFx + jFy .
Since F is conservative, we have therefore,

V V
Fx = , Fy = .
x y
1 2
T=
2
( x + y 2 ) .

1 2
L=
2
( x + y 2 ) V ( x, y ) . . . . (1)
The Hamiltons principle states that

t1
Ldt = 0, . . . (2)
t0
t1
L L L L
x x + y y + x x + y y dt = 0 ,
t0
t1
V V
( x x + y y ) x x y y dt = 0 .
t0
. . . (3)
Consider
t1 t1
d
x xdt
t0
= x
t0
dt
( x )dt

Integrating by parts we get
t1 t1
x xdt
= ( x x )t1
t
x ( x )dt .
0
t0 t0
Since x = 0 at both the ends t0 and t1 along any path, therefore,

t1 t1
x xdt
t0
=
x ( x )dt .
t0
. . . (4)
Similarly, we have
t1 t1
y ydt
t0
=
y ( y )dt .
t0
. . . (5)
Substituting these values in equation (3) we get

t1
V V
t x + x x +

y+
y
y dt = 0 .

0
Since x and y are independent and arbitrary, then we have

V V
x+
= 0, y+
= 0.
x y
V
x=
= Fx ,
x
. . . (6)
V
y=
= Fy .
y
These are the equations of motion of a particle of unit mass moving under the action
of the conservative force field.
Example 2: Use Hamiltons principle to find the equation of motion of a simple
pendulum.
Solution: In case of a simple pendulum, the only generalized co-ordinate is , and
the Lagrangian is given by (Refer Ex. 26 of Chapter I)
1 2 2
L= ml mgl (1 cos ) . . . . (1)
2

The Hamiltons Principle states that the path followed by the pendulum is one along
which the line integral of Lagrangian is extremum. i.e.,
t1
Ldt = 0 ,
t0
t1
1
2 ml mgl (1 cos ) dt = 0 ,
2 2
t0
t1
ml
mgl sin dt = 0 .
2

t0
d d
Since, we have = .
dt dt
Therefore
t1
d
ml dt ( ) mgl sin dt = 0 .
2
t0
Integrating the first integral by parts we get

t1
t1

ml 2 ( ) t0
m l 2 + gl sin dt = 0 .
t0
t
Since ( )t1 = 0 , we have therefore,
0
t1
m l + gl sin dt = 0 .
2
t0
As is arbitrary, we have
l 2 + gl sin = 0
g
+ sin = 0 .
l
This is the required equation of motion of the simple pendulum.

Spherical Pendulum :
Example 3 : A particle of mass m is moving on the surface of the sphere of radius r
in the gravitational field. Use Hamiltons principle to show the equation of motion is
given by
p2 cos g
2 4 3
sin = 0 ,
m r sin r
where p is the constant of angular momentum.
Solution: Let a particle of mass m be moving on the surface of a sphere of radius r.

The particle has two degrees of freedom and hence two generalized co-ordinates
, . The Lagrangian of the motion is (Refer Ex. 28 of Chapter I) given by
1 2 2
L=
2
( )
mr + sin 2 2 mgr cos . . . . (1)
The Hamiltons Principle states that the path followed by the particle between two
time instants t0 and t1 is one along which the line integral of Lagrangian is
extremum. i.e.,
t1
Ldt = 0 ,
t0
t1
1
2 mr ( )
+ sin 2 2 mgr cos dt = 0,
2 2
t0
t1
mr (
+ sin + sin cos 2 ) + mgr sin dt = 0.
2 2

t0
d d
Since, we have = .
dt dt
Therefore,
t1
d d
mr
2
dt
2

2
(
( ) + sin ( ) + mr sin cos + mgr sin dt = 0.
2 )
t0
dt
Integrating the first two integrals by parts we get

t1
) mr 2 g
t1 t1
2
mr ( ) 2
+ mr sin 2
( 2
sin cos sin dt

t0 t0

t0
r
t1
d
mr 2
dt
( )
sin 2 dt = 0.
t0
t1 t
Since ( )t 0
= 0 = ( )t1 ,
0
we have therefore,
t1 t
2 g
1
2 d
(
t mr sin cos r sin dt + t mr dt sin dt = 0
2 2
)
0 0
Since and are independent and arbitrary, hence we have

g
sin cos 2 sin = 0,
r
d
mr 2
dt
( )
sin 2 = 0 mr 2 sin 2 = p ( const.)
Eliminating we obtain
p2 cos g
2 4 3
sin = 0 . . . . (2)
m r sin r
as the required differential equation of motion for spherical pendulum.
Unit 2:
Hamiltonian Formulation :
Introduction:
We have developed Lagrangian formulation as a description of mechanics in
terms of the generalized co-ordinates and generalized velocities with time t as a
parameter in Chapter I and the equations of motion were used to solve some
problems. We now introduce another powerful formulation in which the independent
variables are the generalized co-ordinates and the generalized momenta known as
Hamiltons formulation. This formulation is an alternative to the Lagrangian

formulation but proved to be more convenient and useful, particularly in dealing with
problems of modern physics. Hence all the examples solved in the Chapter I can also
be solved by the Hamiltonian procedure. As an illustration few of them are solved in
this Chapter by Hamiltons procedure.
The Hamiltonian Function:
The quantity p q
j
j j L when expressed in terms of
q1 , q2 ,..., qn , p1 , p2 ,... pn , t is called Hamiltonian and it is denoted by H.
Thus H = H ( q j , p j , t ) = p j q j L .
j
Hamiltons Canonical Equations of Motion :

Theorem 6 : Define the Hamiltonian and hence derive the Hamiltons canonical
equations of motion.
Proof : We know the Hamiltonian H is defined as
H = H ( q j , p j , t ) = p j q j L . . . . (1)
j
Consider H = H (qj , p j ,t ) . . . . (2)
We find from equation (2) that

H H H
dH = dp j + dq j + dt . . . . (3)
j p j j q j t
Now consider H = p j q j L .
j
Similarly we find
dH = q j dp j + dq j p j dL,
j j
L L L
dH = q j dp j + dq j p j dq j dq j dt . . . . (4)
j j j q j j q
j t
We know the generalized momentum is defined as

L
pj = .
q j
Hence equation (4) reduces to

L L
dH = q j dp j dq j dt . . . (5)
j j q j t
Now comparing the coefficients of dp j , dq j and dt in equations (3) and (5) we get
H L H L H
q j = , = , = . . . . (6)
p j q j q j t t
However, from Lagranges equations of motion we have

L
p j =
q j
Hence equations (6) reduce to

H H
q j = , p j = . . . . (7)
p j q j
These are the required Hamiltons canonical equations of motion. These are the set of
2n first order differential equations of motion and replace the n Lagranges second
order equations of motion.
Derivation of Hamiltons equations of motion from Hamiltons Principle :
Theorem 7 : Obtain Hamiltons equations of motion from the Hamiltons principle.
Proof: We know the action of a particle is defined by
t1
I = Ldt . . . (1)
t0
where L is the Lagrangian of the system. If H ( p j , q j , t ) is the Hamiltonian of the
motion then we have by definition

H = p j q j L . . . . (2)
j

Replacing L in equation (1) by using (2) we have the action in mechanics as
t1 t1

I = Ldt = p j q j H dt . . . . (3)
t0 t0 j
Now by Hamiltons principle, we have
t1 t1

Ldt = 0 p j q j H dt = 0 . . . . (4)
t0 t0 j
This is known as the modified Hamiltons principle. Thus we have
t1 t1

Ldt = p j q j H dt ,
t0 t0 j
t1 t1
H H H
Ldt = p j q j + p j q j qj pj t dt.
t0 t0 j j j q j j p j t
Since time is fixed along any path, hence change in time along any path is zero. i.e.,
t = 0 along any path. Hence above equation becomes
t1 t1
H H
Ldt = q j p j + p j q j q j dt . . . (5)
p j j q j
t0 t0 j j
Now consider the integral

t1 t1
d
p q dt = p
t0 j
j j
t0 j
j
dt
( q j ) dt .
Integrating the integral on the r. h. s. by parts we get
t1
t1
t1
t j j j j j j t j p j q j dt .
p q
dt = p q
0 t0 0
t1
Since ( q j )t = 0 . We have therefore
0
t1 t1
p j q j dt = p j q j dt .
t0 j t0 j

t1 t1
H H
Ldt = q j p j + p j + q j dt .
p j q j
t0 t0 j j
Now we see that
t1 t1
H H
Ldt = 0 q j p j + p j + q j dt = 0 .
p j q j
t0 t0 j j
For holonomic system we have q j , p j are independent, hence
t1
H H
Ldt = 0 q j = 0, p j + = 0.
t0
p j q j
t1
H H
Ldt = 0 q j = , p j = . . . . (6)
t0
p j q j
These are the Hamiltons canonical equations of motion.

Remark :
We see from equation (6) that the Hamiltons canonical equations of motion
are the necessary and sufficient conditions for the action to have stationary value.
Example 4 : Show that addition of the total time derivative of any function of the
form f ( q j , t ) to the Lagrangian of a holonomic system, the generalized momentum
and the Jacobi integral are respectively given by

f f
pi + and H .
qi t
Does the new Lagrangian L unchanged the Hamiltons principle? Justify your
claim.
Solution: Let the new Lagrangian function after addition of the time derivative of
any function of the form f ( q j , t ) to the Lagrangian L be denoted by L . Thus we
have

df
L = L + . . . . (1)
dt
Thus the generalized momentum corresponding to the new Lagrangian L is defined
by
L
pj = . . . . (2)
q j
L df
Thus pj = + ,
q j q j dt
f f
pj = p j + qk + ,
q j k qk t
f
pj = p j + . . . . (3)
q j
This is the required generalized momentum corresponding to the new Lagrangian L .

Similarly, the Jacobi integral for new function L is given by
H = pj q j L,
j
L df
H = q j L + .
j q j dt
f df
H = pj + q j L + ,

j q j dt

f
H = ( p j q j L ) ,
j t
f
H = H . . . . (4)
t
This is a required Jacobi integral for the new Lagrangian L .
Now we show that the new Lagrangian L also satisfies the Hamiltons principle.
Therefore, consider

2 2 2
df
Ldt = Ldt + dt ,
1 1 1
dt
2 2 2
Ldt = Ldt + df ,
1 1 1
2 2
Ldt = Ldt + ( f )1 ,
2
1 1
2
2 2
f f
Ldt = Ldt + qj + t .
1 1 j q j t
1
But in variation time is held fixed along any path and hence t = 0 along any
path.
Further, co-ordinates at the end points are held fixed.
2
( q )
j 1 = 0.
Hence we have from the above equation that

2 2
Ldt = Ldt .
1 1
Thus the Hamiltons Principle

2 2
Ldt = 0 Ldt = 0 .
1 1
This shows that the new Lagrangian L satisfies the Hamiltons principle.
Lagrangian from Hamiltonian and conversely :
Example 5: Obtain Lagrangian L from Hamiltonian H and show that it satisfies
Lagranges equations of motion.
Solution: The Hamiltonian H is defined by
H = p j q j L . . . . (1)
j
which satisfies the Hamiltons canonical equations of motion.

H H
q j = , p j = . . . . (2)
p j q j
Now from equation (1) we find the Lagrangian
L = p j q j H . . . . (3)
j
and show that it satisfies Lagranges equations of motion. Thus from equation (3) we
have
L H L
= , and = pj .
q j q j q j
Now consider
L d L H d
= ( pj ),
q j dt q j q j dt
L d L
= p j p j ,
q j dt q j
L d L
= 0.
q j dt q j
This shows that the equation (3) gives the required Lagrangian which satisfies the
Lagranges equations of motion.
Example 6 : Obtain the Hamiltonian H from the Lagrangian and show that it
satisfies the Hamiltons canonical equations of motion.
Solution: The Hamiltonian H in terms of Lagrangian L is defined as
H = p j q j L . . . . (1)
j
where L satisfies the Lagranges equations of motion viz.,
L d L
= 0 , . . . (2)
q j dt q j
L d L
= ,
q j dt q j
d
= ( p j ).
dt

L
= p j . . . . (3)
q j
Now from equation (1) we find

H L
= . . . . (4)
q j q j

H
= p j . . . . (5)
q j
Similarly, we find from equation (1)

H
= q j . . . . (6)
p j
Equations (5) and (6) are the required Hamiltons equations of motion.
Physical Meaning of the Hamiltonian :

Theorem 8 :
1. For conservative scleronomic system the Hamiltonian H represents both a
constant of motion and total energy.
2. For conservative rheonomic system the Hamiltonian H may represent a
constant of motion but does not represent the total energy.
Proof : The Hamiltonian H is defined by
H = p j q j L . . . . (1)
j
where L is the Lagrangian of the system and

L
pj = . . . (2)
q j
is the generalized momentum. This implies from Lagranges equation of motion that

d L L
p j = = . . . (3)
dt q j q j
Differentiating equation (1) w. r. t. time t, we get
dH L L L
= p j q j + p j qj q j qj . . . (4)
dt j j j q j j q
j t
On using equations (2) and (3) in equation (4) we readily obtain

dH L
= . . . (5)
dt t
Now if L does not contain time t explicitly, then from equation (5), we have
dH
=0
dt
This shows that H represents a constant of motion.
However, the condition L does not contain time t explicitly will be satisfied by
neither the kinetic energy nor the potential energy involves time t explicitly.
Now there are two cases that the kinetic energy T does not involve time t explicitly.
1. For the conservative and scleronomic system :
In the case of conservative system when the constraints are scleronomic, the
kinetic energy T is independent of time t and the potential energy V is only function
of co-ordinates. Consequently, the Lagrangian L does not involve time t explicitly
and hence from equation (5) the Hamiltonian H represents a constant of motion.
Further, for scleronomic system, we know the kinetic energy is a homogeneous
quadratic function of generalized velocities.
T = a jk q j qk . . . . (6)
j ,k
Hence by using Eulers theorem for the homogeneous quadratic function of

generalized velocities we have
T
q
j
j
q j
= 2T . . . . (7)

For conservative system we have
L T
pj = = . . . . (8)
q j q j
Using (7) and (8) in the Hamiltonian H we get

H = 2T (T V ) ,
H = T +V = E . . . . (9)
where E is the total energy of the system. Equation (9) shows that for conservative
scleronomic system the Hamiltonian H represents the total energy of the system.
2. For conservative and rheonomic system :

In the case of conservative rheonomic system, the transformation equations
do involve time t explicitly, though some times the kinetic energy may not involve
time t explicitly. Consequently, neither T nor V involves t, and hence L does not
involve t. Hence in such cases the Hamiltonian may represent the constant of motion.
However, in general if the system is conservative and rheonomic, the kinetic energy
is a quadratic function of generalized velocities and is given by
T = a jk q j qk + a j q j + a . . . (10)
j ,k j
where
1 r r
a jk = mi i i ,
i 2 q j qk
. . . (11)
r r
a j = mi i i ,
i q j t
2
1 r
a = mi i .
i 2 t
We see from equation (10) that each term is a homogeneous function of generalized
velocities of degree two, one and zero respectively. On applying Eulers theorem for
the homogeneous function to each term on the right hand side, we readily get

T
q
j
j
q j
= 2T2 + T1 . . . . (12)
where
T2 = a jk q j qk ,
j ,k
T1 = a j q j ,
j
T0 = a
are homogeneous function of generalized velocities of degree two, one and zero
respectively. Substituting equation (12) in the Hamiltonian (1) we obtain
H = T2 T0 + V
showing that the Hamiltonian H does not represent total energy. Thus for the
conservative rheonomic systems H may represent the constant of motion but does not
represent total energy.
Cyclic Co-ordinates In Hamiltonian :
Theorem 9 : Prove that a co-ordinate which is cyclic in the Lagrangian is also cyclic
in the Hamiltonian.
Solution: We know the co-ordinate which is absent in the Lagrangian is called cyclic
L
co-ordinate. Thus if q j is cyclic in L =0.
q j
Hence the Lagranges equation of motion reduces to
d L
= 0 p j = 0 , . . . (1)
dt q j
L
where p j = is the generalized momentum. However, from Hamiltons
q j
canonical equations of motion we have

H
p j = . . . . (2)
q j

Equations (1) and (2) gives
H
=0. . . . (3)
q j
This shows that the co-ordinate q j is also absent in the Hamiltonian, and
consequently, it is also cyclic in H. Thus a co-ordinate which is cyclic in the

Lagrangian is also cyclic in the Hamiltonian.
Worked Examples
Example 7 : Describe the motion of a particle of mass m moving near the surface of
the Earth under the Earths constant gravitational field by Hamiltons procedure.
Solution: Consider a particle of mass m moving near the surface of the Earth under
the Earths constant gravitational field. Let (x, y, z) be the Cartesian co-ordinates of
the projectile, z being vertical. Then the Lagrangian of the projectile is given by
1
L= m ( x 2 + y 2 + z 2 ) mgz . . . . (1)
2
We see that the generalized co-ordinates x and y are absent in the Lagrangian,
hence they are the cyclic co-ordinates. This implies that any change in these co-
ordinates can not affect the Lagrangian. This implies that the corresponding
generalized momentum is conserved. In this case the generalized momentum is the
linear momentum and is conserved.
px = mx = const.
i.e., . . . (2)
p y = my = const.
pz = mz .
This shows that the horizontal components of momentum are conserved.
The Hamiltonian of the particle is defined by
H = p j q j L,
j

1
H = px x + p y y + pz z m ( x 2 + y 2 + z 2 ) + mgz . . . . (3)
2
Eliminating x, y , z between equations (2) and (3) we get
1
H=
2m
( px2 + p y2 + pz2 ) + mgz . . . . (4)
The Hamiltons equations of motion give

H H H
p x = = 0, p y = = 0, p z = = mg . . . . (5)
x y z
H px H p y H pz
and x = = , y = = , z = = . . . . (6)
px m p y m pz m
From these set of equations we obtain

x = 0,
y = 0,
z = g . . . (7)
These are the required equations of motion of the projectile near the surface of the
Earth.
Example 8 : Obtain the Hamiltonian H and the Hamiltons equations of motion of a
simple pendulum. Prove or disprove that H represents the constant of motion and
total energy.
Solution: The Example is solved earlier by various methods. The Lagrangian of the
pendulum is given by
1 2 2
L= ml mgl (1 cos ) , . . . (1)
2
where the generalized momentum is given by
L p
p = = ml 2 = 2 . . . . (2)
ml
The Hamiltonian of the system is given by
H = p L,
1
H = p ml 2 2 + mgl (1 cos ) .
2

Eliminating we obtain
p2
H= + mgl (1 cos ) . . . . (3)
2ml 2
Hamiltons canonical equations of motion are
H H
q j = , p j = .
p j q j
These equations give

p
= , p = mgl sin . . . . (4)
ml 2
Now eliminating p from these equations we get
g
+ sin = 0. . . . (5)
l
Now we claim that H represents the constant of motion.
Thus differentiating equation (3) with respect to t we get
dH p p
= + mgl sin ,
dt ml 2
+ mgl sin ,
= ml 2
g
= ml 2 + sin ,
l
dH
=0.
dt
This proves that H is a constant of motion. Now to see whether H represents total
energy or not, we consider
1 2 2
T +V = ml + mgl (1 cos ) .
2

Using equation (4) we eliminate from the above equation, we obtain
p2
T +V = + mgl (1 cos ) . . . . (6)
2ml 2
This is as same as the Hamiltonian H from equation (3). Thus Hamiltonian H
represents the total energy of the pendulum.
Example 9: The Lagrangian for a particle moving on a surface of a sphere of radius r
is given by
1 2 2
L=
2
( )
mr + sin 2 2 mgr cos .
Find the Hamiltonian H and show that it is constant of motion. Prove or disprove that
H represents the total energy. Is the energy of the particle constant? Justify your
claim.
Solution: We are given the Lagrangian of a particle moving on the surface of a
sphere (Spherical Pendulum) in the form
1 2 2
L=
2
( )
mr + sin 2 2 mgr cos . . . . (1)
We see that is a cyclic co-ordinate. This implies the corresponding generalized

momentum is conserved. i.e.
L
p = = mr 2 sin 2 = const. . . . (2)

L
Similarly, p = = mr 2. . . . (3)

The Hamiltonian of the particle is defined as
1
( )
H = p + p mr 2 2 + sin 2 2 + mgr cos .
2
. . . (4)
Using equations (2) and (3) we obtain the Hamiltonian function
1 p2 p2
H= 2+ 2 2 + mgr cos . . . . (5)
2 mr mr sin

The Hamiltons canonical equations of motion give
cos p2
p = + mgr sin ,
mr 2 sin 3
p = 0,
p . . . (6)
= ,
mr 2
p
= .
mr 2 sin 2
Eliminating p from equation (6) we get the equation of motion of spherical
pendulum as

2
cos p2
mr 2 3 mgr sin = 0 . . . . (7)
mr sin
(i) Now we claim that H is a constant of motion, differentiate equation (5) with
respect to t, we get
dH p p p p p2 cos
= 2
+ 2 2
2 3
mgr sin ,
dt mr mr sin mr sin
Putting the values of p , p from equation (6) we get
p cos p p cos p 2
2
dH p
= 2 2 3 + mgr sin 2 4 3 2 mgr sin
dt mr mr sin m r sin mr

dH
= 0,
dt
showing that H is a constant of motion.
(ii) Now consider the sum of the kinetic and potential energy of the spherical
pendulum, where
1 2 2
T=
2
(
mr + sin 2 2 , )
V = mgr cos

Thus
1 2 2
T +V =
2
( )
mr + sin 2 2 + mgr cos . . . . (8)
We eliminate , from equation (8) by using equation (6) to get
1 p2 p2
T + V = 2 + 2 2 + mgr cos . . . . (9)
2 mr mr sin
We see from equations (5) and (9) that the total energy of the spherical pendulum is
the Hamiltonian of motion. Now to see it is constant or not, multiply equation (7) by
we get
p2 cos
mr 2 mgr sin = 0
mr 2 sin 3
d 1 2 2 d p2 d
mr + 2 2 + ( mgr cos ) = 0.
dt 2 dt 2mr sin dt
Integrating we get
1 2 2 p2
mr
+ 2 2 + ( mgr cos ) = const.
2 2mr sin
Eliminating on using equation (6) we get
1 p2 p2
+ + mgr cos = const. . . . . (10)
2 mr 2 mr 2 sin 2
We see from equations (5), (9) and (10) that the Hamiltonian H represents the total
energy and the energy of the particle is conserved.
Example 10: Two mass points of mass m1 and m2 are connected by a string passing
suspended. Assuming m2 moves only in a vertical line, write down the Hamiltonian
for the system and hence the equations of motion. Prove or disprove that
i) Hamiltonian H represents the constant of motion.
ii) H represents total energy of the system.
Solution: The example is solved in Chapter I. (please refer to Example 24). The
Lagrangian of the system is given by
1 1
L=
2
( )
m1 r 2 + r 2 2 + m2 r 2 + m2 g ( l r )
2
. . . (1)
We see that the co-ordinate is cyclic in the Lagrangian L and hence the
corresponding generalized momentum is conserved.
L
p = = m1r 2 = const. . . . (2)

Similarly, we find
L
pr = = ( m1 + m2 ) r = const. . . . (3)
r
Now the Hamiltonian function is defined as
L L
H = r + L,
r
1 1
H= ( m1 + m2 ) r2 + m1r 22 m2 g ( l r ) .
2 2
Eliminating r and we obtain
pr2 p2
H= + 2 m2 g ( l r ) . . . . (4)
2 ( m1 + m2 ) 2m1r
The Hamilton canonical equations of motion viz.,

H H
p j = , q j =
q j p j
give
H p2 H
p r = = 3 m2 g , p = = 0. . . . (5)
r m1r
H pr H p
r = = , = = 2 . . . . (6)
pr ( m1 + m2 ) p m1r

p2
( m1 + m2 ) r + m2 g = 0 . . . . (7)
m1r 3
This is the required equation of motion.
i) To prove H represents a constant of motion, we differentiate equation (4)
with respect to t. Thus we have
dH pr p r p p p 2 r
= + 2 3 + m2 gr
dt ( m1 + m2 ) m1r m1r
Using equations (5) and (6), we have

dH pr p2 pr m2 g p2 pr m2 gpr
= + ,
dt ( m1 + m2 ) m1r m1 + m2 ( m1 + m2 ) m1r ( m1 + m2 ) r
3 3
dH
= 0.
dt
This shows that The Hamiltonian H represents a constant of motion.
ii) We have the kinetic and potential energies of the system are respectively
given by
1 1
T=
2
( 2
)
m1 r 2 + r 2 2 + m2 r 2 ,
V = m2 g ( l r ) .
Now consider
1 1
T +V =
2
( )
m1 r 2 + r 2 2 + m2 r 2 m2 g ( l r ) .
2
. . . (8)
Eliminating r and from equation (8) on using equations (6) we obtain

pr2 p2
T +V = + 2 m2 g ( l r ) . . . (9)
2 ( m1 + m2 ) 2m1r
From equations (4) and (9) we see that the total energy is equal to the
Hamiltonian function. Thus Hamiltonian H represents total energy of the system. To
prove that the total energy is conserved, multiply the equation of motion (7) by r ,
we get

p2 r
( m1 + m2 )
rr + m2 gr = 0 .
m1r 3
This we write as
d r 2 d p2 d
( m + m ) + + ( m2 gr ) = 0 .
dt 2 dt 2m1r 2 dt
1 2
Integrating and then eliminating r we get

pr2 p2
+ 2 + m2 gr = const. . . . . (10)
2 ( m1 + m2 ) 2m1r
Equations (9) and (10) show that the total energy of the system is conserved.
Note : Equation (10) is the first integral of equation of motion. Its physical
significance is that the Hamiltonian H represents the constant of total energy.
Example 11: A particle of mass m is moving on a xy plane which is rotating about z
axis with angular velocity . The Lagrangian is given by
1
m ( x y ) + ( y + x ) V ( x, y ) .
2 2
L=
2
Show that the Hamiltonian H is given by
1
H=
2m
( px2 + p 2y ) + px y p y x + V .
Find the equations of motion and hence prove or disprove that
i) H represents a constant of motion and
ii) H represents the total energy.
Solution: The Lagrangian of the particle is given by
1
m ( x y ) + ( y + x ) V ( x, y ) .
2 2
L= . . . (1)
2
where the generalized momentum px and p y are given by
L p
px = = m ( x y ) x = x + y , . . . (2)
x m

L py
py = = m ( y x ) y = x . . . . (3)
y m
The Hamiltonian H is defined by H = xp
x + yp
y L
1
H = px x + p y y m ( x y ) + ( y + x ) + V ( x, y ) .
2 2
2
Using equations (2) and (3) we eliminate x and y from the above equation to get
the Hamiltonian of the system
1
H=
2m
( px2 + p 2y ) + ( px y p y x ) + V . . . . (4)
The Hamiltons canonical equations of motions give

H V H V
p x = = p y , p y = = p x ,
x x y y
. . . (5)
H px H p y
x = = + y, y = = x.
px m px m
Solving these equations we obtain the equations which describe the motion
V
m (
x 2 y 2 x ) = ,
x
. . . (6)
V
m (
y + 2 x 2 y ) =
y
Now to prove whether H is a constant of motion or not, differentiate equation (4)
w. r. t. t to get
dH 1 V V
= ( px p x + p y p y ) + ( p x y + px y p y x p y yx ) + x + y .
dt m x y
Using (5) we have
dH
=0.
dt
This shows that H represents the constant of motion. Now to show whether H
represents total energy or not, we have the total energy of the system
E = T +V ,

1
E=
2m
( px2 + p y2 ) + V ( x, y ) . . . . (7)
We see from equations (4) and (7) that the Hamiltonian H does not represent total
energy.
Example 12 : A bead slides on a wire in the shape of a cycloid described by
equations
x = a ( sin ) , y = a (1 + cos ) , 0 2 .
Find the Hamiltonian H, hence the equations of motion. Also prove or disprove that
i) H represents a constant of motion
ii) H represents a total energy.
Solution: A particle describes a cycloid whose equations are
x = a ( sin ) , y = a (1 + cos ) , 0 2 . . . . (1)

1
T= m ( x 2 + y 2 ) ,
2
where
x = a (1 cos ) ,
y = a sin .
Hence the kinetic energy of the particle becomes
T = ma 2 2 (1 cos ) ,
The potential energy of the particle is given by

V = mga (1 + cos ) .

L = ma 2 2 (1 cos ) mga (1 + cos ) . . . . (2)
The Hamiltonian H of the particle is

L
H = L, . . . (3)

where from equation (2) we have
L
= p = 2ma 2 (1 cos ) . . . . (4)

Using equations (2) and (4) in (3) we obtain the expression for Hamiltonian as
H = ma 2 2 (1 cos ) + mga (1 + cos ) . . . . (5)
Using equation (4) we eliminate from equation (5) to get the required Hamiltonian
H as
p2
H= + mga (1 + cos ) . . . . (6)
4ma 2 (1 cos )
The Hamiltons canonical equations of motion are

H p2 sin
p = = + mga sin , . . . (7)
4ma (1 cos ) 2
2
H p
= = . . . . (8)
p 2ma (1 cos )
2
From equations (7) and (8) we obtain the equation of motion of the particle
p2 sin g
(1 cos ) + 2 4 3
sin = 0 . . . (9)
8m a (1 cos ) 2a
Eliminating p from equation (9) we obtain the equation which describes the motion
of the particle in the form
2ma 2 (1 cos ) + ma 2 sin 2 mga sin = 0 . . . . (10)
Now to prove
i) H is a constant of motion, differentiate equation (6) with respect to time t we
get
dH 2 p p p2 sin
= mga sin .
dt 4ma 2 (1 cos ) 4ma 2 (1 cos )2
Using equations (7) and (8) we readily get

dH
=0.
dt
This shows that the Hamiltonian H is a constant of motion.
ii) H represents the total energy
We find from the expressions for kinetic energy and potential energy that
T + V = ma 2 2 (1 cos ) + mga (1 + cos ) . . . . (11)
Eliminating from equation (11) we get equation (6) that gives the required
expression for the Hamiltonian. Now multiply equation (10) by we get
2ma 2 (1 cos )
+ ma 2 sin 3 mga sin = 0 .
This can be written as

d
ma 2 (1 cos ) 2 + mga (1 + cos ) = 0 .
dt
Integrating we get
H = T + V = ma 2 2 (1 cos ) + mga (1 + cos ) = const.
This shows that the Hamiltonian H represents the constant of total energy.
Example 13 : Obtain the Hamiltons equation of motion for a one dimensional
harmonic oscillator.
Solution: The one dimensional harmonic oscillator consists of
a mass attached to one end of a spring and other end of the
spring is fixed. If the spring is pressed and released then on
account of the elastic property of the spring, the spring exerts a
force F on the body in the opposite direction. This is called F M
x
restoring force. It is found that this force is proportional
to the displacement of the body from its equilibrium position.
Fx
F = kx
where k is the spring constant and negative sign indicates the force is opposite to the
displacement. Hence the potential energy of the particle is given by
V = Fdx,
V = kxdx + c,
kx 2
V= + c,
2
where c is the constant of integration. By choosing the horizontal plane passing
through the position of equilibrium as the reference level, then V=0 at x=0. This
gives c=0. Hence potential energy of the particle is
1 2
V= kx . . . . (1)
2
The kinetic energy of the one dimensional harmonic oscillator is
1 2
T= mx . . . . (2)
2
Hence the Lagrangian of the system is
1 2 1 2
L= mx kx . . . . (3)
2 2
The Lagranges equation motion gives
k
x + 2 x = 0, 2 =
. . . . (4)
m
This is the equation of motion. is the frequency of oscillation.
The Hamiltonian H of the oscillator is defined as
H = xp
x L,
1 1
x mx 2 + kx 2 ,
H = xp
2 2
where
L p
px = = mx x = x .
x m
Substituting this in the above equation we get the Hamiltonian
px2 1 2
H= + kx . . . . (5)
2m 2

Solving the Hamiltons canonical equations of motion we readily get the equation (4)
as the equation of motion.
Example 14: For a particle the kinetic energy and potential energy is given by
respectively,
1 2 1 r 2
T= mr , U = 1 + 2 .
2 r c
Find the Hamiltonian H and determine

1. whether H = T + V
dH
2. whether =0.
dt
Solution: The kinetic and potential energies of a particle are given by
1 2 1 r 2
T= mr , U = 1 + 2
2 r c
respectively. The Lagrangian function is therefore given by
1 2 1 r 2
L= mr 1 + 2 . . . . (1)
2 r c
We see that the particle has only one degree of freedom and hence it has only one
generalized co-ordinate. The generalized momentum is defined by
L 2r
pr = = mr 2 ,
r rc
pr r c 2
r = . . . . (2)
( mrc2 2 )
Thus the corresponding Hamiltonian function is defined by
H = pr r L,
1 1 r 2 . . . (3)
H = pr r mr 2 + 1 + 2 .
2 r c
Eliminating r between (2) and (3) we obtain the Hamiltonian H as

1 pr2 r c 2 1
H= + . . . . (4)
2 ( mrc 2 ) r
2
1. Now the sum of the kinetic and potential energies is given by
1 2 1 r 2
T +U = mr + 1 + 2 . . . . (5)
2 r c
Eliminating r between (2) and (5) we get
1 pr r c ( mrc + 2 ) 1
2 2 2
T +U = + . . . . (6)
2 ( mrc 2 2 )2 r
We see from equations (4) and (6) that the Hamiltonian H does not represent the total
energy.
T +U H .
2. Now differentiating equation (4) w. r. t. time t we get
dH p p rc 2 pr2 rc
2 r
= r 2r , . . . (7)
dt ( mrc 2 ) ( mrc 2 2 )2 r2
where
2 2r 2
p r =
r m 2 + 2 2 .
rc r c
Substituting this in equation (7) and simplifying we get
dH r3 r
= pr r + 2 2 2 .
dt r c r
dH
0.
dt
This shows that the Hamiltonian H is not a constant of motion.
Example 15 : A particle is thrown horizontally from the top of a building of height h
with an initial velocity u. Write down the Hamiltonian of the problem. Show that H
represents both a constant of motion and the total energy.

Solution: Let the particle be thrown horizontally from the top of a building of
y
height h with an initial velocity u. The motion
of the particle is in a plane. If P (x, y) are the
position co-ordinates of the particle at any
P(x, y) instant, then its kinetic energy and the potential
h energy are respectively given by

y
x
O
1
T= m ( x 2 + y 2 ) , . . . (1)
2
V = mg ( h y ) . . . . (2)

1
L= m ( x 2 + y 2 ) + mg ( h y ) . . . . (3)
2
The particle has two degrees of freedom and hence two generalized co-ordinates. We
see that the generalized co-ordinate x is cyclic in L, hence the corresponding
generalized momentum is conserved.
L
px = = mx = const.
x
. . . (4)
L
py = = my.
y
The Hamiltonian function H is defined as
H = p j q j L
j
1
H = px x + p y y m ( x 2 + y 2 ) mh ( h y ) . . . (5)
2
Eliminating the velocities from equations (4) and (5) we obtain the Hamiltonian of
motion as

1
H=
2m
( px2 + p 2y ) mg ( h y ) . . . . (6)
The corresponding Hamiltons canonical equations of motion are

H px H p y
x = = , y = = ,
px m p y m
H H
and p x = = 0, p y = = mg.
x y
Solving these equations we get the equations of motion as
x = 0,
y = g . . . . (7)
Now differentiating equation (6) with respect to t we get
dH 1
= ( px p x + p y p y ) + mgy
dt m
dH
= 0,
dt
This proves that H is a constant of motion.
Now to see whether H represents total energy or not, we consider
1
T +V = m ( x 2 + y 2 ) mg ( h y )
2
Putting the values of x and y we obtain the expression for the Hamiltonian as
1
T +V =
2m
( px2 + p y2 ) mg ( h y ) . . . . (8)
This represents the Hamiltonian H, proving that H represents the total energy of the
particle.
Example 16: A particle is constrained to move on the arc of a parabola x 2 = 4ay

under the action of gravity. Show that the Hamiltonian of the system is
2a 2 px2 mg 2
H= + x .
m ( 4a + x ) 4 a
2 2

Is the Hamiltonian of the particle representing total energy? Is it a constant of
motion?
Solution: Given that a particle is constrained to move on the arc of the parabola
x 2 = 4ay . . . (1)
where y is vertical axis, under the action of gravity. The kinetic energy of the particle
is given by
1
T= m ( x 2 + y 2 ) . . . (2)
2
and the potential energy is given by
V = mgy . . . . (3)
However, x and y are not the generalized co-ordinates, because they are related by
the constraint equation (1). Eliminating y from equations (2) and (3) on using (1) we
obtain
1 2 x2 x2
T= mx 1 + 2 , V = mg .
2 4a 4a
Hence the Lagrangian function is
1 2 x2 x2
L= mx 1 + 2 mg . . . . (4)
2 4a 4a
Now we see that the system has one degree of freedom and only one generalized co-
ordinate x.
L x2
px = = mx 1 + 2 ,
x 4a
4a 2 px
x = . . . . (5)
m ( 4a 2 + x 2 )
Now the Hamiltonian H becomes

H = xp
x L

1 2 x2 x2
H = xp
x - mx 1 + 2 + mg . . . . (6)
2 4a 4a
On using (5) we write equation (6) as
2a 2 px2 mgx 2
H= + . . . . (7)
m ( 4a 2 + x 2 ) 4a
This is the required Hamiltonian function. Now to see whether this H represents total
energy or not, we consider
1 2 x2 x2
T+V = mx 1 + 2 +
mg . . . . (8)
2 4 a 4a
Using equation (5) we obtain
2a 2 px2 mgx 2
T +V = + . . . . (9)
m ( 4a 2 + x 2 ) 4a
Which is the Hamiltonian of the motion, showing that it represent the total energy of
the particle. Now to show that the Hamiltonian H represents constant of motion, we
first find the equation of motion. From equation (4) we have
L m mgx
= 2 xx 2 ,
x 4a 2a
L x2
= 1 + 2 mx.
x 4a
Hence the equation of motion becomes
d L L d x2 m mgx
=0 2
1 + 2 mx 2 xx + = 0,
dt x x dt 4a 4a 2a
( 4a 2
+ x 2 )
x + xx 2 + 2agx = 0 . . . . (10)
Now differentiating equation (7) with respect to t we get

x2
dH 4a 2 px p x xxp + mg xx .
=
dt m ( 4 a + x ) ( 4a 2 + x 2 ) 2 a
2 2 2

Eliminating px , p x we obtain
dH m
= 2 ( 4a 2 + x 2 )
x + xx 2 + 2agx x .
dt 4a
dH
This implies from equation (10) that =0.
dt
This shows that the Hamiltonian H is a constant of motion.
Example 17 : Set up the Hamiltonian for the Lagrangian
m 2 2
L ( q, q , t ) = q sin t + qq sin 2t + q 2 2 .
2
Derive the Hamiltons equations of motion. Reduce the equations in to a single
second order differential equation.
Solution: The Lagrangian of the system is given by
m 2 2
L ( q, q , t ) = q sin t + qq sin 2t + q 2 2 . . . (1)
2
The system has only one degree of freedom and hence only one generalized co-
ordinate q. The generalized momentum is given by
L m
p= = ( 2q sin 2 t + q sin 2t ) . . . (2)
q 2
1 p q
q = 2 sin 2t . . . . (3)
sin t m 2
Now the Hamiltonian function H is defined as
m 2 2
H = pq
2
( q sin t + qq sin 2t + q 2 2 ) . . . . (4)
Substituting the value of q from equation (3) in (4) and simplifying we get
p2 q 2 m 2 m
H= 2
pq cot t + cos 2 t q 2 2 . . . . (5)
2m sin t 2 2
This is the Hamiltonian of the system. The Hamiltons canonical equations of motion
give

H p
q = = q cot t . . . . (6)
p m sin 2 t
and p = p cot t q 2 m cos 2 t + mq 2 . . . . (7)

From equation (6) we find
m
p= 2q sin 2 t + q sin 2t . . . (8)
2
Differentiating equation (8) w. r. t. t we get
m
p = 2q sin 2 t + 4q sin t cos t + q sin 2t + 2q 2 cos 2t . . . (9)
2
Equating equations (7) and (9) we get
q + 2 q cot t 2q 2 = 0 . . . . (10)
This equation determines the motion of the particle.
Example 18 : A Lagrangian of a system is given by
m k
L ( x, y, x, y ) =
2
( + cy 2 ) ( ax 2 + 2bxy + cy 2 ) ,
ax 2 + 2bxy
2
where a, b, c, k , m are constants and b 2 ac 0 . Find the Hamiltonian and
equations of motion. Examine the particular cases a = 0, c = 0 and b = 0, c = a .
Solution: Given that
m k
L ( x, y, x, y ) =
2
( + cy 2 ) ( ax 2 + 2bxy + cy 2 ) ,
ax 2 + 2bxy
2
. . . (1)
where a, b, c, k , m are constants and b 2 ac 0 . We see that the system has two
generalized co-ordinates x and y. Hence the corresponding generalized momenta are
L
px = = m ( ax + by ) , . . . (2)
x
L
and py = = m ( bx + cy ) . . . . (3)
y
Solving these equations for x and y we get

cpx bp y bpx ap y
x = , y = . . . . (4)
m ( b ac )
2
m ( b 2 ac )
The Hamiltonian H is defined by

H = p j q j L,
j
m k
H = px x + p y y
2
( + cy 2 ) + ( ax 2 + 2bxy + cy 2 ) . ...(5)
ax 2 + 2bxy
2
Using equations (4) in (5) we obtain after simplifying
1 a c k
H= bpx p y p y2 px2 + ( ax + 2bxy + cy 2 ) .. . . (6)
m ( b ac )
2
2 2 2
This is the required Hamiltonian of the system. The Hamiltons equations of motion
corresponding to two generalized co-ordinates x, y are
H
p x = = k ( ax + by ) ,
x
. . . (7)
H
p y = = k ( bx + cy ) .
y
and
H 1
x = =
px m ( b ac )
2 ( bp y cpx )
. . . (8)
H 1
y = =
p y m ( b ac )
2 ( bpx cp y )
From equations (2), (3) and (7) we have
m ( ax + by) + k ( ax + by ) = 0,
. . . (9)
m ( bx + cy) + k ( bx + cy ) = 0.
These are the required equations of motion. Solving these equations for x and y we
obtain respectively
mx + kx = 0. . . . (10)
my + ky = 0. . . . (11)

The solutions of these equations are
k k
x = c1 cos t + c2 sin t , . . . (12)
m m
k k
and y = d1 cos t + d 2 sin t. . . . (13)
m m
Now the cases a = 0, c = 0 and b = 0, c = a yield from equations (9) the same set
of equations (10) and (11).
Example 19 : The Lagrangian for a system can be written as
y
L = ax 2 + b + fy 2 xz
+ czy + gy k x 2 + y 2 ,
x
where a, b, c, f, g and k are constants. What is Hamiltonian? What quantities are
conserved ?
Solution: The Lagrangian of the system is
y
L = ax 2 + b + fy 2 xz
+ czy + gy k x 2 + y 2 , . . . (1)
x
where a, b, c, f, g and k are constants. The system has three degrees of freedom and
has three generalized co-ordinates (x, y, z), of which z is cyclic. This implies the
corresponding generalized momentum pz is conserved.
L
pz = = cy + fy 2 x = const. . . . (2)
z
Similarly, we find
L
px = = 2ax + fy 2 z , . . . (3)
x
L b
and py = = + cz + g . . . . (4)
y x
Solving these equations for x, y , z we get

1 fy 2 b
x = p x py g ,
2a c x
1 fy 2 fy 2 b
y = pz x
p py g ,
c 2a c x
1 b
z = p y g . . . . (5)
c x
The Hamiltonian of the system is defined as
H = px x + p y y + pz z L
y
H = px x+ z z ax 2 b
p y y+p fy 2 xz
cxy gy + k x 2 + y 2 . . . . (6)
x
The required Hamiltonian is obtained by eliminating x, y , z from equation (6).
Unit 3: Rouths Procedure :

Introduction:
The presence of cyclic co-ordinates in the Lagrangian L is not much
profitable because even if the co-ordinate q j does not appear in L, the corresponding
generalized momentum q j generally does, so that one has to deal the problem with all
variables and the system has n degrees of freedom. However, if q j is cyclic in the
Hamiltonian then p j is constant and then one has to deal with the problem involving
only 2n-2 variables, i.e., only n-1 degrees of freedom. Hence Hamiltonian procedure
is especially adapted to the problems involving cyclic co-ordinates. The advantage of
Hamiltonian formulation in handling with cyclic co-ordinates is utilized by Routh
and devised a method by combining with the Lagrangian procedure and the method
is known as Rouths Procedure. The Method is described in the following theorem.

Theorem 10: Describe the Rouths procedure to solve the problem involving cyclic
and non-cyclic co-ordinates.
Proof: Consider a system of particles involving both cyclic and non-cyclic co-
ordinates. Let q1 , q2 , q3 ,..., qs of q1 , q2 , q3 ,..., qn are cyclic co-ordinates, then a new
function R, known as the Routhian is defined as
s
R ( q1 , q2 ,..., qn ; p1 , p2 ,..., ps ; qs +1 , qs + 2 ,..., qn , t ) = p j q j L ( q j , q j , t ) . . . (1)
j =1
The Routhian R is obtained by modifying the Lagrangian L so that it is no longer a

function of the generalized velocities corresponding to the cyclic co-ordinates, but
instead involves only its conjugate momentum. The advantage in doing so is that p j
can then be considered one of the constants of integration and the remaining
integrations involve only the non-cyclic co-ordinates.
Now we take R = R ( q1 , q2 ,..., qn ; p1 , p2 ,..., ps ; qs +1 , qs + 2 ,..., qn , t ) , and find the total
differential dR as
n s n
R R R R
dR = dq j + dp j + q dq j + dt . . . . (2)
j =1 q j j =1 p j j = s +1 j t
Now we consider
s
R = p j q j L ( q j , q j , t )
j =1
and find the total differential as

s s
dR = p j dq j + q j dp j dL,
j =1 j =1
s s n n
L L L
dR = p j dq j + q j dp j dq j dq j dt.
j =1 j =1 j =1 q j j =1 q
j t

s s s L n
L
dR = p j dq j + q j dp j dq j + q dq
j =1 q j
j =1 j =1 j j = s +1 j
s L n
L L
dq j + q dq dt.
j =1 q j
t
j j = s +1 j
s s n
L L
dR = q j dp j dq j dq j
j =1 j =1 q j j = s +1 q j
n
L L
q dq
j = s +1
j
t
dt. . . . (3)
j
Now equating the corresponding coefficients on both the sides of equations (2) and
(3) we obtain
R
= q j , j = 1, 2,..., s . . . (4)
p j
R L
= = p j , j = 1, 2,..., s . . . (5)
q j q j
R L
= = p j , j = s + 1, s + 2,..., n . . . (6)
q j q j
R L
and = = pj, j = s + 1, s + 2,..., n . . . (7)
q j q j
We see that for cyclic co-ordinates q1 , q2 ,..., qs equations (4) and (5) represent
Hamiltons equations of motion with R as the Hamiltonian, while equations (6) and
(7) for the non-cyclic co-ordinates q j ( j = s + 1, s + 2,..., n ) represent Lagranges
equations of motion with R as the Lagrangian function. i.e., from equations (6) and
(7) we obtain
d R R
= 0, j = s + 1, s + 2,..., n . . . (8)
dt q j q j
Thus by Routhian procedure a problem involving cyclic and non-cyclic co-ordinates
can be solved by solving Lagranges equations for non-cyclic co-ordinates with

Routhian R as the Lagrangian function and solving Hamiltonian equations for the
given cyclic co-ordinates with R as the Hamiltonian function. In this way The
Routhian has a dual character Hamiltonian H and the Lagrangian L.
Worked Examples
Example 20 : Find Lagrangian L, Hamiltonian H and the Routhian R in spherical

polar co-ordinates for a particle moving in space under the action of conservative
force.
Solution: Let a particle be moving in a space. If (x, y, z) are the Cartesian co-
ordinates and ( r , , ) are the spherical co-ordinates of the particle, then we have the
relation between them as

x = r sin cos ,
y = r sin sin , . . . (1)
z = r cos .
1
The kinetic energy T = m ( x 2 + y 2 + z 2 ) of the particle, in spherical polar co-
2
ordinates becomes
1
T=
2
( )
m r 2 + r 2 2 + r 2 sin 2 2 . . . . (2)
Since the force is conservative, hence the potential energy of the particle is the
function of position only.
V = V ( r , , ) . . . . (3)
Hence the Lagrangian function of the particle becomes

1
L=
2
( )
m r 2 + r 2 2 + r 2 sin 2 2 V ( r , , ) . . . . (4)
We see that is cyclic in L, hence the corresponding generalized momentum is

conserved. i.e.,

L
p = = mr 2 sin 2 = const. . . . (5)

Similarly we find
L
pr = = mr,
r
. . . (6)
L
p = = mr 2.

H = p j q j L,
j
. . . (7)
1
( )
H = pr r + p + p m r 2 + r 2 2 + r 2 sin 2 2 + V .
2
Eliminating the generalized velocities r,, between equations (5), (6) and (7) we
get
1 2 1 2 1 2
H= pr + 2 p + 2 2 p + V . . . . (8)
2m r r sin
Now the Routhian R is defined by
R = p L, . . . (9)
This becomes after eliminating r,, between (5), (6) and (9) we get
p2 1
( )
R r , , , r, , t = 2 ( )
m r 2 + r 2 2 + V .
2mr sin 2 2
. . . (10)
Example 21 : A planet moves under the inverse square law of attractive force, Find
Lagrangian L, Hamiltonian H, and the Routhian R for the planet.
Solution: A motion of a planet is a motion in the plane. If ( r , ) are the generalized
co-ordinates of the planet then its kinetic and potential energies are respectively
given by
1 K
T=
2
( )
m r 2 + r 2 2 , V = .
r
Hence the Lagrangian function is defined by
1 K
L=
2
(
m r 2 + r 2 2 + .
r
) . . . (1)
We see that is the cyclic co-ordinate in L. This implies that the corresponding
angular momentum of the planet is conserved.
L p
p = = mr 2 = const. = 2 . . . . (2)
mr
L p
Also pr = = mr r = r . . . (3)
r m
H = p j q j L,
j
1 K
(
H = pr r + p m r 2 + r 2 2
2 r
)
On using equations (2) and (3) we obtain
1 2 p2 K
H= pr + 2 . . . . (4)
2m r r
This is the required Hamiltonian.

Now the Routhian is defined as
R ( r , , p , r, t ) = p L
1 K
( )
R ( r , , p , r, t ) = p m r 2 + r 2 2 .
2 r
Eliminating we get
p2 1 K
R ( r , , p , r, t ) = 2
mr 2 . . . . (5)
2mr 2 r
This can also be written as
p2 pr2 K
R ( r , , p , r, t ) = . . . . (6)
2mr 2 2m r

Principle of Least Action :
Action in Mechanics :
In Mechanics the time integral of twice the kinetic energy is called the action.
Thus
t1
A = 2Tdt
t0
is called the action.

t1
i.e. A = p j q j dt
t0 j
is called action in Mechanics.
Principle of Least Action :

There is another variational principle associated with the Hamiltonian
formulation and is known as the principle of least action. It involves a new type of
variation which we call the - variation.
In - variation the co-ordinates of the end points remain fixed while the time is
allowed to vary. The varied paths may terminate at different points, but still position
co-ordinates are held fixed.
Mathematically, we have
I dI
I = d , I = d .
d
Thus for the family of paths represented by the equation
q j = q j ( , t ) , t = t ( )
We have
dq j q dt
q j = d = j + q j d .
d d

q j dt
q j = d + q j d
d .
q j = q j + q j t
This shows that the total variation is the sum of two variations.
Worked Examples
Example 22 : If f = f ( q j , q j , t ) then show that
df
f = f + t .
dt
Solution: Consider a system of particles moving from one point to another. Let the
family of paths between these two points be given by
q j = q j ( t , ) . . . . (1)
In variation time is not held fixed, it depends on the path. This implies that
t = t ( ) . . . (2)
Since f = f ( q j , q j , t ) . . . (3)
then we find variation in f as
f f f
f = q j + q j + t . . . . (4)

j q j q j t

However, we have q j = q j + q j t . . . (5)
Similarly we find
q j = q j + qj t , . . . (6)

f f f
f = ( q j + q j t ) + ( q j + qj t ) + t .

j q j q j t


f f f f f f
f = qj + q j + t + q j + qj + t

j q j q j t j q j q j t
Note here that the term t added because it is zero, since in variation time t is
held fixed and consequently change in time t is zero. This can be written as
df
f = f + t . . . . (7)
dt
Since f is arbitrary, we can write it as
d
= + t . . . . (8)
dt
Theorem 11 : For a conservative system for which the Hamiltonian H is conserved,

the principle of least action states that
t1
p j q j dt = 0 .
t0 j
Proof: Consider a conservative system for which the Hamiltonian H is conserved.

Let AB be the actual path and CD be the varied path. In - variation the end points
of the two paths are not terminated at the same point. The end points A and B after
t take the positions C and D such that the position co-ordinates of A, C and B, D
are held fixed. Now we know the action is given by
y t1
A = p j q j dt
t0 j
B D
t1
A = ( L + H )dt
t0
t1
A = Ldt + H ( t )t1 ,
t
0
A C t0
t0 t0 + t0 x
O t1 t1 + t1

t1
A = Ldt + H ( t1 t0 ) . . . . (1)
t0
Thus
t1
A = Ldt + H ( t )t1 .
t
. . . (2)
0
t0
Since time limits are also subject to change in -variation, therefore cant be
taken inside the integral. Let
t1
Ldt = I
t0
I = L .
Therefore
I = I + It .
Thus we have
t1 t1
Ldt = Ldt + L ( t )t1 .

t
0
t0 t0
t1 t1
L L L
Ldt =
t
qj + q j + t dt + L ( t )t1 .

j q j
t0
q j t
0
t0
Since in variation, time is held fixed along any path, hence there is no variation in
time, therefore change in time is zero. Thus we have
t1
L
t1
L
Ldt =
t
qj + q j dt + L ( t )t1 .
q j q j
t0 j
0
t0
Using Lagranges equations of motion we write this equation as

t1 t1
Ldt = ( p j q j + p j q j )dt + L ( t )t1 .

t
0
t0 t0 j
Since
dq j d
= qj .
dt dt

Hence we have
t t

1 1
d
Ldt = p j q j + p j q j dt + L ( t )t1 .
t
t0 j
t0
dt 0
d
t1 t1
Ldt = ( p j q j ) dt + L ( t )t0 .
t1

t0 t0
dt j
Since
d
= + t
dt
Hence above integral becomes
t1 t1
d
Ldt = d p j t q j dt + L ( t )t1 .
t
t0 t0 j dt 0
t1 t1
t1

Ldt = p j q j p j q j t + L ( t )t1
t
j t0 j t0
0
t0
Since in variation, position co-ordinates at the end points are fixed.

t1
( q ) j t
0
= 0.
Consequently above equation reduces to

t1
t1

Ldt = ( p j q j L ) t
t0 j t0
t1
Ldt = ( H t )t1
t
0
t0

A = 0,
t1
.
i.e., p j q j dt = 0
t0 j
Thus the system moves in space such that -variation of the line integral of twice
the kinetic energy is zero. This proves the principle of least action.

Example 23 : A system of two degrees of freedom is described by the Hamiltonian
H = q1 p1 q2 p2 aq12 + bq22 , a, b are const.
p1 aq1 p2 bq2
Show that i) , ii ) , iii ) q1q2 iv) H are constant of motion.
q2 q1
Solution: The Hamiltonian of a dynamical system is given by
H = q1 p1 q2 p2 aq12 + bq22 , a, b are const. . . . (1)
where we see that q1 , q2 are the generalized co-ordinates. The Hamiltons canonical
equations of motion are
H
p j = p1 = 2aq1 p1 ,
q j . . . (2)
p 2 = p2 2bq2 ,
and
H
q j = q1 = q1
p j . . . (3)
q2 = q2 .
Now to show
p1 aq1
1) is a constant of motion, consider
q2
d p1 aq1 q2 ( p1 aq1 ) ( p1 aq1 ) q2

= .
dt q2 q22
Using equations (2) and (3) we obtain
d p1 aq1 p1 aq1
=0 = const.
dt q2 q2
Similarly we prove that

p1 aq1 p2 bq2
= const., = const., q1q2 = const.
q2 q1

Now to prove the Hamiltonian H is also constant, we differentiate equation (1) with
respect to t to get
dH
= q1 p1 + q1 p1 q2 p2 q2 p 2 2aq1q1 + 2bq2 q2 .
dt
Using equations (2) and (3) we see that
dH
= 0 H = const. .
dt
This shows that H is a constant of motion.
Example 24 : A Lagrangian for a particle of charge q moving in the electromagnetic
field of force is given by
1 2
L= mv + q ( v A ) q .
2
Find the Hamiltonian H, the generalized momenta.
Solution: The Lagrangian of a particle moving in the electromagnetic field is given
by
1 2
L= mv + q ( v A ) q . . . . (1)
2
We write this expression as
1
L= m ( x 2 + y 2 + z 2 ) + q ( xA
x + yA z ) q .
y + zA . . . (2)
2
where is a scalar potential function of co-ordinates only. We see that x, y, z are
the generalized co-ordinates. Hence the corresponding generalized momenta become
L
pj = px = mx + qAx ,
q j
p y = my + qAy ,
pz = mz + qAz .
Solving these equations for velocity components we get

1
x = ( px qAx ) ,
m
1
y = ( p y qAy ) , . . . (3)
m
1
z = ( pz qAz ) .
m
The Hamiltonian of the particle is given by
H = p j q j L,
j
1
H = xp
x + yp z m ( x 2 + y 2 + z 2 ) q ( xA
y + zp x + yA z ) + q . . . (4)
y + zA
2
Eliminating x, y , z from equation (4) by using equation (3) we get
1 q 1 2 2
H= ( px2 + p y2 + pz2 ) ( px Ax + p y Ay + pz Az ) + q ( Ax + Ay2 + Az2 ) + q . . (5)
2m m 2m
This can be written in vector notions as
1 2
H=
2m
( p qA ) + q . . . . (6)
This is the required Hamiltonian of the particle moving in the electromagnetic field.
H
The Hamiltons equation of motion q j = gives the same set of equations (3),
p j
H
while the equation p j = gives
q j
H q q2
p x = =
x m x
( p A
x x + p A
y y + p z z)
A
2m x
( Ax2 + Ay2 + Az2 ) q
x
.

p x = q
x
( v A) q
x
.
Similarly, other two components are given by

p y = q
y
( v A) q ,
y

p z = q ( v A ) q .
z z
All these three equations can be put in to the single equation as
p = q + q ( v A ) . . . . (7)
Exercise:
1. The Lagrangian of an anharmonic oscillator of unit mass is
1 2 1 2 2
L= x x x3 + xx , , are constants.
2 2
Find the Hamiltonian and the equation of motion. Show also that
(i) H is a constant of motion and
(ii) H T +V .
1 2 1
Ans : H = ( px x ) + 2 x 2 + x 3 .
2 2
x + 2 x + 3 x 2 = 0 .
Equation of motion
2. Find the Hamiltonian and the equations of motion for a particle constrained to
move on the surface obtained by revolving the line x = z about z axis. Does
it represent the constant of motion and the constant of total energy?
Hint: Surface of revolution is a cone x 2 + y 2 = z 2
pr2 p2
Ans.: H = + + mgr .
4m 2mr 2
p2 g
r
2 3
+ = 0, p = mr 2 a const. of motion .
2m r 2

3. Let a particle be moving in a field of force given by
1 r 2 2rr
F = 2 1 .
r c2
Find the Hamiltonian H and show that it represents the constant of motion
and also total energy.
Ans. : Refer Example (25) of Chapter I; the potential energy of the particle is
given by
1 r 2
V = 1 + 2 .
r c
pr2 1
The Hamiltonian becomes H = + .
2 r
2 m 2
rc
4. A sphere of radius a and mass m rests on the top of a fixed trough sphere of
radius b. The first sphere is slightly displaced so that it rolls without
slipping. Obtain the Hamiltonian of the system and hence the equation of
motion. Also prove that H represents a constant of motion and also total
energy.
7 2
Ans. : H = m ( a + b ) 2 + mg ( a + b ) cos .
10
5. A particle is constrained to move on the plane curve xy = c , c is a constant,
under gravity. Obtain the Hamiltonian H and the equations of motion. Prove
that the Hamiltonian H represents the constant of motion and total energy.
Ans. : Refer Example (20) of Chapter I for the Lagrangian L and is given by
1 2 c 2 mgc
L= mx 1 + 4 .
2 x x
px2 mgc
The Hamiltonian H becomes H = + .
c 2
x
2m 1 + 4
x

6. A body of mass m is thrown up an inclined plane which is moving
horizontally with constant velocity v. Use Hamiltons procedure to find the
equations of motion. Prove that the Hamiltonian H represents the constant of
motion but does not represent the total energy.
Ans. : For the Lagrangian function, refer Example (28) of Chapter I. The
Hamiltonian of motion is
pr2 1
H= pr v cos + mgr sin mv 2 sin 2 .
2m 2
7. A particle moves on the surface characterized by
x = r cos , y = r sin , z = r cot .
Find the Hamiltonian H and prove that it represents the constant of motion
and also the constant of total energy.
2
p 2 sin 2 p
Ans. : H = r + 2 + mgr cot .
2m 2mr
r r sin 2 2 + g cos sin = 0 .
The equation of motion is
8. Find the Hamiltonian and the Hamiltons canonical equations of motion for
the Lagrangian given by
1 1
( ) 2
( ) 2
2
L r , r, , = m r 2 + r 2 2 + mgr cos k ( r r0 ) ,
where k , m, g , r0 are constants.
pr2 p2 1 2
Ans : H = + 2 mgr cos + k ( r r0 )
2m 2mr 2
Equations of motion:
mr mr 2 mg cos + k ( r r0 ) = 0,
2 g
+ r + sin = 0.
r r

CHAPTER - IV
THE KINEMATICS OF RIGID BODY
Unit 1: Rigid Body:

Introduction:
In this chapter we define a rigid body and describe how the number of
degrees of freedom of a rigid body with N particles is determined. There are two
types of motion involved in the case of rigid body viz.; the translation and the
rotation. Various sets of variables have been used to describe the orientation of rigid
body. We will discuss in this chapter how the Eulerian angles and the complex
Cayley-Klein parameters can be used for the description of rigid body with one point
fixed.
Geometrically, matrix represents rotation; we will find the matrix of
transformation in terms of Eulerian angles and Cayley-Klein parameters and
establish the relation between them. This unit is devoted to the study of orthogonal
transformations and its properties.
Rigid Body :
A rigid body is regarded as a system of many (at least three) non-collinear
particles whose positions relative to one another remain fixed. i.e., distance between
any two of them remains constant through out the motion. The internal forces
holding the particles at fixed distances from one another are known as forces of
constraint. These forces of constraint obey the Newtons third law of motion.

Worked Examples
Example 1 : Explain how the generalized co-ordinates of a rigid body with N

particles reduce to six for its description.
Solution :
Generalized co-ordinates of rigid body :
A system of N particles free from constraints can have 3N degrees of freedom
and hence 3N generalized co-ordinates. But the constraints involved in rigid body
with N particles are holonomic and scleronomic and are given by
rij = aij , i j = 1, 2,...., N . . . (1)
where rij denotes the distance between the i th and j th particles, and aij are constants.
Equation (1) is symmetric in i and j and i j as the distance of the i th from itself is
zero, therefore, the possible number of constraints is
N N ( N 1)
C2 = . . . . (2)
2
N ( N 1)
We notice that for N > 7, > 3 N . Therefore the actual number of
2
degrees of freedom cannot be obtained simply by subtracting the number of
constraints from 3N. This is simply because all constraints in equation (1) are not
independent.
To show how the generalized co-ordinates of a rigid
C
body with N particles reduce to six for its
P
description, let a rigid body be regarded as a system
A of at least three non-collinear particles whose
positions relative to one another remain fixed. Thus a
B
system of 3 particles free from constraints has 9 degrees of freedom but there
involves 3 constraints. Hence the number of generalized co-ordinates reduces to six.
Thus the total number of degrees of freedom for three non-collinear particles A, B,
and C of a rigid body is equal to six. This is because each particle has 3-degrees of
freedom and less three equations of constraints.
The position of each further particle say P requires three more co-ordinates
for its description, but there will be three equations of constraints for this particle,
because the distance of P from A, B, C is fixed. Thus three co-ordinates for P and
less three equations of constraints for P gives zero degrees of freedom. Thus any
other particle apart from A, B, C taken to specify the configuration of the rigid body
will not add any degrees of freedom. Once the positions of three of the particles of
the rigid body are determined the constraints fix the positions of all remaining
particles.
Thus the configuration of the rigid body would be completely specified by
only three particles i.e., by six degrees of freedom, no matter how many particles it
may contain.
Example 2 : Describe the motion of the rigid body.

Solution: A rigid body can have two types of motion
(i) a translational motion and
(ii) a rotational motion.
Thus a rigid body in motion can be completely specified if its position and
orientation are known.
However, if one of the points of a rigid body is fixed, the translation motion
of the body is absent and the body rotates about any line through the fixed point.
Again, if we fix up a second point, then the motion of the body is restricted to rotate
about the line joining the two fixed points. Further, if we also fix the third point of
the body non-collinear with other two, the position of the body is fixed and there is
no motion of any kind. The co-ordinates of the third point alone will be able to locate
the rigid body completely in space. It follows that the position of the rigid body is
determined by any three non-collinear points of it that is by six degrees of freedom.

Of the six generalized co-ordinates, 3 co-ordinates are used to describe translational
motion and other three co-ordinates are used to describe rotational motion. Since a
rigid body with one point fixed has no translational motion and hence it has 3-
degrees of freedom and three generalized co-ordinates-which are used to describe the
rotational motion.
Orthogonal Transformation :
Example 3 : Define orthogonal transformation. Show that finite rotation of a rigid
body about a fixed point of the body is not commutative.
Solution : Consider ( x1 , x2 , x3 ) and ( x1, x2 , x3 ) be two co-ordinate systems. The
general linear transformation between these two co-ordinate systems is defined by

the following set of equations
x1 = a11 x1 + a12 x2 + a13 x3 ,
x2 = a21 x1 + a22 x2 + a23 x3 , . . . (1)
x3 = a31 x1 + a32 x2 + a33 x3 ,
where a11 , a12 ,..., a33 are constants. These three equations can be combined in to a
single equation as
3
xi = aij x j , i = 1, 2,3. . . . (2)
j =1
where A = ( aij ) is called the matrix of transformation. Let a vector
r = x1i + x2 j + x3 k defined in ( x1 , x2 , x3 ) co-ordinate system be transformed to
( x1, x2 , x3 ) co-ordinate systems in the form r = x1i + x2 j + x3 k . Since the magnitude
of the vector must be the same in both the co-ordinate system, we must have
therefore
3 3
x = x
i =1
i
2
i =1
2
i . . . . (3)

Using equation (2) in equation (3) we get
3
3
3
3
a x a ij j ik xk = xi2 ,
i =1
i =1 j =1 k =1
3 3
3 3
ij ik j k xi .
j =1 k =1 i =1
a a

x x =
i =1
2
Equating the corresponding coefficients on both the sides of the above equation we
get
3
a a
i =1
ij ik = jk , . . . (4)
where jk is the Kronecker delta symbol and is defined by
jk = 0 when j k,
. . . (5)
= 1 when j = k.
Thus any transformation (2) satisfying (4) is called as an orthogonal transformation.
Ex. 4. Show that two successive finite rotations of a rigid body about a fixed point of
the body are not commutative.
Solution: Consider two successive linear transformations described by the
matrices B and A corresponding to two successive displacements of the rigid body.
Let the first transformation from x to x be denoted by the matrix B and is defined
by
3
xk = bkj x j , k = 1, 2,3, . . . (6)
j =1
where the matrix of transformation is B = ( bkj ) .
Let the succeeding transformation from x to x be defined by the matrix

A = ( aik ) and is given by
3
xi = ai k xk , i = 1, 2, 3. . . . (7)
k =1

Now the transformation from x to x is obtained by combining the two equations
(6) and (7) as
3
xi = a
j , k =1
b x j , i = 1, 2,3.
i k kj
This may also be written as

3
xi = cij x j , i = 1, 2,3. . . . (8)
j =1
where C = ( cij ) is the matrix of transformation from x to x and the elements of the
matrix of transformation are defined as

3
cij = aik bkj . . . . (9)
k =1
These elements are obtained by multiplying the two matrices A and B. Thus the two
successive linear transformations described by A and B is equivalent to a third linear
transformation described by the matrix C, defined by
C = AB. . . . (10)
Since the matrix multiplication is not commutative in general, hence the finite
rotations of a rigid body about a fixed point of the body are not commutative.
Properties of orthogonal transformation matrix :

Example 5: Prove that the product of two orthogonal transformations is again
orthogonal transformation.
Solution: Consider two successive orthogonal linear transformations of a rigid body
with one point fixed corresponding to two successive displacement of the rigid body
and are described by the matrices B and A respectively. We know that the two
successive orthogonal transformations is equivalent to a third linear transformation
described by the matrix C, where C = AB, and its elements are defined by

cij = aik bkj . . . (1)
k
where A = ( aik ) , B = ( bkj ) are matrices of orthogonal transformations.
This implies that
a a
i
ij ik = jk , . . . (2)
b b
i
ij ik = jk . . . . (3)
Consider now

c c
i
ij ik = aimbmj ail blk
i m l
= aimbmj ail blk ,
i , m ,l

= aim ail bmj blk ,
m ,l i
= ml bmj blk ,
m ,l
= blj blk ,
l
c c
i
ij ik = jk .
This proves that the product of two orthogonal transformations is again an

orthogonal transformation.
Note: Though the matrix multiplication is not commutative in general, but it is
associative when the product is defined.
i.e., ( AB ) C = A ( BC ) .
Example 6: Show that in the case of an orthogonal transformation the inverse matrix
is identified by the transpose of the matrix.
Solution: Consider an orthogonal transformation from xi to xi described by the
matrix A and is given by
3
xi = aij x j , i = 1, 2,3. . . . (1)
j =1
where A = ( aij ) is the matrix of transformation satisfying the condition
a a
i
ij ik = jk . . . . (2)
Let the matrix of inverse transformation from xi to xi be described by the inverse
matrix A1 = ( aij ) , aij are the elements of the inverse matrix of transformation
satisfying
a a
i
ij ik = jk . . . . (3)
Also we have
AA1 = I a i
a = kj .
ki ij . . . (4)
Now consider the double sum a

k ,i
kl aki aij which can be evaluated either by summing
over k first or over i first. Therefore we evaluate the double sum as

a
k ,i
kl aki aij = akl aki aij ,
i k
= li aij ,
i
a
k ,i
kl aki aij = alj . . . . (5)
Now evaluating the double sum over i first and then over k , we obtain

a
k ,i
kl aki aij = aki aij akl ,
k i
= kj akl ,
k
a
k ,i
kl aki aij = a jl . . . . (6)

aij = a ji , i, j ,
( aij ) = ( a ji ) ,
A1 = A . . . . (7)
This proves that the inverse matrix of an orthogonal transformation identifies the
transpose matrix.
Example 7 : Show that the determinant of an orthogonal matrix is 1 .

Solution : Let A be the matrix of an orthogonal transformation and A1 be its
inverse matrix. Then we have
AA1 = I . . . (1)
However, we know that in the case of an orthogonal matrix, its inverse is identified
by its transpose.
A1 = A .
AA = I . . . (2)
Taking the determinant on both the sides of above equation we get
AA = 1
A A = 1,
2
A = 1 as A = A ,
A = 1.
Infinitesimal Rotation :
Example 8: Define infinitesimal rotation. Show that infinitesimal rotation of a rigid
body with one point fixed is commutative. Also find the inverse matrix of
infinitesimal rotation.

Solution: An infinitesimal rotation is an orthogonal transformation of co-ordinate
axes in which the components of a vector are almost the same in both the sets of
axes. The new co-ordinates differ from the old co-ordinates by an infinitesimal
amounts.
Mathematically, an infinitesimal transformation is defined as
x1 = x1 + 11 x1 + 12 x2 + 13 x3 ,
x2 = x2 + 21 x1 + 22 x2 + 23 x3 ,
x3 = x3 + 31 x1 + 32 x2 + 33 x3 ,
i.e., xi = xi + ij x j , . . . (1)
where summation is defined over the repeated index j and ij are the elements of the
matrix of infinitesimal transformation and are infinitesimal. i.e., second order terms
in ij can be neglected. Hence we write equations (1) as
xi = ij x j + ij x j ,
i.e., xi = ( ij + ij ) x j , . . . (2)
In matrix notations we write equation (2) as

X = (I + ) X , . . . (3)
x1 x1

where X = x2 , X = x2 , I = ( ij ) , = ( ij )
x x
3 3
and I + is the matrix of infinitesimal transformation.
Now let I + 1 and I + 2 be two matrices of successive infinitesimal
transformations.
Consider
( I + 1 )( I + 2 ) = I .I + I 2 + 1I + 1 2
= I + 1 + 2 .

Similarly consider
( I + 2 )( I + 1 ) = I .I + I 1 + 2 I + 21
= I + 1 + 2 .
We see from above equations that
( I + 1 )( I + 2 ) = ( I + 2 )( I + 1 )
This shows that the product of the matrices of two successive infinitesimal
transformations is commutative. Now to find the inverse matrix of an infinitesimal
transformation, consider
( I + )( I ) = I I + I
( I + )( I ) = I
This shows that the inverse matrix of an infinitesimal transformation is I .
1
i.e., (I + ) = (I ) .
For orthogonal transformation we know that its transpose matrix identifies the
inverse matrix. Hence we have
1
(I + ) = ( I + ) = ( I ) ,
where is the transpose of . Consequently we have = .

This shows that the matrix of infinite transformation is anti-symmetric.
Example 9: A constant vector X is given by X = i + 4 j + 2 3k with respect to a

particular co-ordinate system. Find the form of the vector with respect to co-ordinate

system obtained from the first by rotating it about the x-axis through an angle in
3
the anti-clock wise direction. Determine its magnitude and compare with X .

Solution: The matrix of rotation about x-axis through an angle in the anti clock
3
wise direction is given by

1 0 0 1 0 0

sin = 0
1 3
A = 0 cos .
3 3 3 2 2

0 sin
cos 0 3 1

3 3 2 2
Hence the new vector with respect to the new co-ordinate axes obtained from the first

by rotating through an angle about x-axis is given by
3

1 0 0
1 1

X = 0
1 3
4 = 5
2 2 2 3 3

3 1
0
2 2
X = i + 5 j 3k
The magnitude of this new vector is X = 29 .
This shows that X = X .
Eulers Theorem 1 : Show that the general displacement of a rigid body with one
point fixed is a rotation about some axis passing through the fixed point.
Proof : Consider a rigid body with one point fixed and be taken as the origin of the
body set of axes. Then the displacement of the rigid body involves no translation of
the body axes, the only change is in orientation. Hence the body set of axes at any
time t can always be obtained by a single rotation of the space set of axes.
Thus any vector lying along the axis of rotation must have the same
components in both the initial and final axes. Further, the orthogonality condition

implies that the magnitude of a vector parallel to the axis of rotation is unaffected. It
means that the vector R has same components in both the system.
R = AR = R . . . . (1)
This is the special case of more general equation
R = AR = R . . . . (2)
where is called as the eigen (characteristic) value. Equation (2) can be written as
( A I ) R = 0 . . . . (3)
The equation (3) can have solution only when

A I = 0 . . . . (4)
This is known as characteristic equation, and the roots of the equation (4) are known
as characteristic values. Since the matrix of rotation A is orthogonal, and then we
have
A1 = A . . . (5)
where A is the transpose of A . This orthogonal matrix satisfies the equation
A = A = 1 . . . . (6)
Now to prove the Eulers Theorem, we just prove that eigen value = 1 . Thus
consider the expression
( A I ) A = AA IA
= AA1 A
( A I ) A = I A . . . . (7)
Taking the determinant of equation (7) we get
( A I ) A = I A
(A I) A = I A
A I = A I . . . (8)
A I = 0. . . . (9)
Comparing equations (4) and (9) we get = 1 . This proves the theorem.
Unit 2: Eulerian Angles :
We have seen that a rigid body with one point fixed has three degrees of
freedom and hence three generalized co-ordinates. To describe the orientation of a
rigid body about a fixed point we use a matrix of rotation, whose elements are called
the direction cosines, which are not linearly independent, therefore they are not
suitable as generalized co-ordinates. So we cannot use them in the description of
Lagrangian of the system. Therefore three new independent parameters are necessary
for the description of a rigid body with one point fixed. A number of such sets of
parameters have been used in the literature but the most common and found to be
useful is the set of Eulerian angles.
Euler has designed three independent parameters called as Eulerian angles, to
describe the orientation of a rigid body with one point fixed. These can be used to
write Lagrangian and hence the Lagranges equations of motion. We shall define the
Eulerian angles and show how these angles can be used for the description of the
orientation of the rigid body.
Theorem 2 : Define Eulerian angles. Obtain the matrix of transformation from space
co-ordinates to body co-ordinates in terms of Eulerian angles. Prove further that this
matrix is orthogonal and hence deduce the matrix of inverse transformation from the
body set of axes to space set of axes.
z = z1
Proof : Eulerian angles , , are the three
successive angles of rotation about a specified
axes performed in specific sequence. These
y1
angles can be used as generalized co-ordinates to

fix the orientation of a rigid body with one point
y
O fixed. Thus the orientation of a rotating body

x
with one point fixed can be completely specified
x1 by three independent Eulerian angles.

To discuss the rotation of the rigid body, let one of the points of the body be
fixed. This implies that there is no translational motion but the body rotates about an
axis passing through the fixed point. Consider two co-ordinate systems, one of which
is (x, y, z) fixed in space (called an inertial frame) and the other ( x, y, z ) fixed in
the body called the body set of axes (also known as non-inertial frame). It has been
observed that the configuration of the rigid body is completely specified by locating
the body set of axes relative to the co-ordinate axes fixed in space. This is achieved
by finding the matrix of transformation from the space set of axes to the body set of
axes. Therefore we shall carry out the transformation from space set of axes to body
set of axes such that x, y, z coincides with x, y, z . This is achieved by three
successive rotations about specified axes.
The sequence will be started by rotating the initial system of axes x, y, z
through an angle anti-clock wise direction about z axis. Let the resulting co-
ordinate system be labeled as x1 , y1 , z1 axes as shown in the fig. In this case xy plane
becomes x1 y1 plane. The rotation is affected by the following transformation

equations.
x1 = x cos + y sin ,
y1 = x sin + y cos , . . . (1)
z1 = z.
These equations can be written in matrix form as
x1 x

y1 = D y or X 1 = DX
z
1 z
where
cox sin 0

D = sin cos 0 . . . (2)
0 0 1

is the matrix of transformation.

z = z1
z2 The second rotation is performed about the new
x1 axis. The axes x1 , y1 , z1 are rotated about x1
y2 axis counter clockwise direction by an angle .

y1 Let the resultant set of co-ordinate

axes be relabeled as x2 , y2 , z2 . Here x2 axis
y
O
being the line of intersection of xy plane and
x x1 y1 plane is called the line of nodes. The
x1 = x2
transformation equations from x1 , y1 , z1 to new
set of axes x2 , y2 , z2 can be represented by the following set of equations:
x2 = x1 ,
y2 = y1 cos + z1 sin , . . . (3)
z1 = y1 sin + z1 cos ,
x2 x1

i.e., y2 = C y1 or X 2 = CX 1
z z
2 1
where
1 0 0

C = 0 cos sin . . . (4)
0 sin cos

z = z1
z2 = z3= z
Finally, the third rotation is performed about
z2 axis. The x2 , y2 , z2 axes are rotated counter y3 = y
y2
clockwise direction by an angle about z2 axis to
y1

produce the third and the final set of axes x3 , y3 , z3 ,
y
O
which coincide, with body set of axes x, y, z . This

x
x1 = x2
x3 = x
completes the transformation from space set of axes to body set of axes. This
transformation is represented by
x3 = x = x2 cos + y2 sin ,
y3 = y = x2 sin + y2 cos , . . . (5)
z3 = z = z 2
or
x x2

y = B y2
z z
2
or X = BX 2
where
cox sin 0

B = sin cos 0 . . . (6)
0 0 1

Thus the space set of axes x, y, z coincides with body set of axes through
three successive rotations , , , which are described by matrices D, C and B. The
angles , , are called Eulerian angles. The Eulerian angles completely specify the
orientation of the x, y, z system relative to the x, y, z system. Now we can obtain
the complete matrix of transformation from x, y, z to x, y, z by writing the matrix
as the triple product of the separate rotations.
X = BX 2
= B ( CX 1 )
= BC ( X 1 )
X = BCDX
X = AX ,

x x

y = A y . . . (7)
z z

where A = BCD is the complete matrix of transformation from x, y, z to x, y, z and
is the product of the successive matrices. Using the equations (2), (4) and (6) the
matrix of transformation from space co-ordinates to the body co-ordinates is then
given by
cos sin 01 0 0 cos sin 0

A = sin cos 0 0 cos sin sin cos 0
0 0 1 0 sin cos 0 0 1

cos cos cos sin sin , cos sin + sin cos cos , sin sin

A = sin cos cos cos sin , sin sin + cos cos cos , cos sin .
sin sin , sin cos , cos

...(8)
This is the required matrix of transformation.
We will now show that this matrix A is orthogonal. Let the matrix A be
represented by
A = ( aij ) .
The condition for orthogonal transformation is that
a a
i
ij ik = jk ,
where ij is a Kronecker delta symbol.
i.e., a1 j a1k + a2 j a2 k + a3 j a3k = jk .
We take the case j = k = 1 , and consider

2
a112 + a21
2
+ a312 = ( cos cos cos sin sin ) +
2 2
+ ( sin cos cos cos sin ) + ( sin sin )
a112 + a21
2 2
+ a31 = 1.
Similarly, we can show for all value of j = k.
Now for j k ,j, k=1, 2, 3 we consider the case
a11a12 + a21a22 + a31a32 =
= ( cos cos cos sin sin )( cos sin + cos sin cos ) +
+ ( sin cos cos cos sin )( sin sin + cos cos cos ) + ( sin sin )( sin cos )
a11a12 + a21a22 + a31a32 = 0 .
Similarly, we can show for all j k . that
a a
i
ij jk =0.
Hence the matrix A is orthogonal. To find the inverse of A, we know that in the case
of orthogonal matrix A1 is the same as the transpose of A. Thus we have
cos cos cos sin sin , cos sin cos cos sin , sin sin
1
A = sin cos + cos cos sin , sin sin + cos cos cos , cos sin

The matrix A1 is the desired matrix, which gives the inverse transformation from
the body set of axes to the space set of axes.
This completes the answer.
Worked Examples
Example 10 : If the matrix of transformation from space set of axes to body set of
axes is equivalent to a rotation through an angle about some axis through the
origin then show that

+
cos = cos cos .
2 2 2
Solution : We know the matrix of complete rotation from space set of axes to body
set of axes in terms of Eulerian angles , , is given by
cos cos cos sin sin , cos sin + sin cos cos , sin sin

A = sin cos cos cos sin , sin sin + cos cos cos , cos sin .

. . . (1)
It is given that this matrix of rotation is equivalent to the matrix of rotation of
co-ordinate axes through an angle about some axis with the same origin.
Equivalently, it means that it is always possible by means of some similar
transformation, to transform the matrix A to the matrix B obtained by rotating the co-
ordinate axes through an angle about some axis with the same origin. This matrix
of rotation is given by either
cos sin 0 1 0 0

B = sin cos 0 or B = 0 cos sin . . . . (2)
0 0 1 0 sin cos

It is well known that under similar transformation trace of the matrix is invariant.
Using this result we have
Trace of A = Trace of B
cos cos cos sin sin + cos cos cos sin sin + cos = 2 cos + 1
2 cos + 1 = ( cos cos sin sin ) + cos ( cos cos sin sin ) + cos
= [1 + cos ] cos ( + ) + cos
+ 2 2
2 cos + 1 = 2 cos 2 2 cos 2 1 + cos sin
2 2 2 2

+
2 cos + 1 = 4 cos 2 cos 2 1
2 2
+
2 ( cos + 1) = 4 cos 2 cos 2
2 2
+
cos 2 = cos 2 cos 2
2 2 2
+
cos = cos cos . . . . (3)
2 2 2
Moments of Inertia and Products of Inertia :

Moment of Inertia :
A uniformly rotating body possesses the tendency to oppose any change in its
state of rotation motion. This quantity is called the moment of inertia.
Theorem 3 : Obtain the angular momentum of a rigid body about a fixed point of the
body when the body rotates instantaneously with angular velocity in terms of
inertia tensor.
Proof: Consider a rigid body composed of N particles having masses mi and position
vectors ri with respect to the fixed point of the body. Since the translational motion
is absent and the body rotates about an axis passes through the fixed point. Let be
the instantaneous angular velocity of the body. If vi is the linear velocity of the
i th particle, then we know it is given by

vi = ri . . . . (1)
If L is the total angular momentum then it is equal to the sum of the angular
momenta of an individual particle.
Therefore we have
N
L = I i = ri mi vi .
i i =1

N
L = ri mi ( ri ) ,
i =1
N
L = mi ri ( ri ) .
i =1
Using the vector identity

a ( b c ) = ( a .c ) b ( a .b ) c
we obtain
N
L = mi ri 2 mi ( ri . ) ri . . . . (2)
i =1 i
Let the components of the position vector ri , the angular velocity and the angular
momentum vector L be denoted by
ri = ixi + jyi + kzi ,
= i x + j y + k z , . . . (3)
L = iLx + jLy + kLz .
Using equations (3) we write

iLx + jLy + kLz =
m ( i
i
i x + j y + k z ) ri 2 ( xix + yi y + zi z ) ( ixi + jyi + kzi )

iLx + jLy + kLz = i x mi ( ri 2 xi2 ) y mi xi yi z mi xi zi +
i i i

i.e., + j x mi xi yi + y mi ( ri 2 yi2 ) z mi yi zi +
i i i

+ k x mi xi zi y mi yi zi + z mi ( ri 2 zi2 )
i i i
Equating the corresponding coefficients on both the sides of the equation we get
Lx = x mi ( ri 2 xi2 ) y mi xi yi z mi xi zi , . . . (4a)
i i i

Ly = x mi xi yi + y mi ( ri 2 yi2 ) z mi yi zi , . . . (4b)
i i i
Lz = x mi xi zi y mi yi zi + z mi ( ri 2 zi2 ) . . . . (4c)
i i i
We write the components of angular momentum as

Lx = I xx x + I xy y + I xz z ,
Ly = I yx x + I yy y + I yz z , . . . (5)
Lz = I zx x + I zy y + I zz z ,
where the coefficients of x , y , z are respectively defined by
I xx = mi ( ri 2 xi2 ) = mi ( yi2 + zi2 ),

i i
I yy = mi ( ri y
2 2
i ) = m (x i
2
i + zi2 ), . . . (6)
i i
I zz = mi ( ri 2 zi2 ) = mi ( xi2 + yi2 ).

i i
and are called the moment of inertia about x, y, and z axes respectively. Equations
(6) show that the moment of inertia is the sum over the particles in the system of the
product of masses and the square of its perpendicular distance from the axis of
rotation. Also the quantities I xy , I xz and I yz are called the product of inertia and are
defined by
I xy = mi xi yi = I yx ,
i
I xz = mi xi zi = I zx , . . . (7)
i
I yz = mi yi zi = I zy .
i
Now the equation (5) can be written in the matrix form as
Lx I xx I xy I xz x

Ly = I yx I yy I yz y . . . (8)
L I I zy I zz z
z zx

or L = I . . . (9)

where I is called the moment of inertia tensor or inertia tensor. Note that moment
of inertia tensor is symmetric and hence it has only six independent components. The
moment of inertia tensor depends only on the mass distribution in the body. It is
I I xy I xz
xx
given by I = I yx I yy I yz . . . . (10)
I I zy I zz
zx
The diagonal elements are called the moments of inertia of the body about the
given point and the given set of body set of axes. The off diagonal components of
moment of inertia tensor are called the product of inertia of the body about the given
point and the given set of body axes.
Note that it is always possible to find a set of axes with respect to which all
the products of inertia tensor vanish leaving off diagonal terms and the axis is called
Principal axis of the body.
Kinetic Energy of a rigid body with one point fixed :

Theorem 4 : Find the kinetic energy of a rigid body rotating about a fixed point of
the body when the moments and products of inertia of the body relative to the set of
axes through fixed point are known.
Proof: Consider a rigid body composed of N particles having masses mi and rotating
with instantaneous angular velocity . If one of the points of the rigid body is fixed
then the translational motion is absent and the body rotates about an axis passes
through the fixed point. If vi is the linear velocity of the i th particle and position
vector ri with respect to the fixed point, then we know it is given by
vi = ri . . . . (1)
We know the kinetic energy of the body is given by
1
T= mi vi2 .
2 i
. . . (2)

Using equation (1) we write (2) as
1
T= mi vi ( ri ) .
2 i
On using the vector identity
a.(b c ) = b .( c a ) ,
we write the expression for the kinetic energy as

1
T= mi. ( ri vi ),
2 i
1
= ( ri mi vi ),
2 i
1
= ri pi
2 i
1
T = .L . . . (3)
2
where L = ri pi
i
is the total angular momentum. Now to express kinetic energy in terms of moment of
inertia and product of inertia, we know the angular momentum of the rigid body is
given by

L = I . . . (4)

where I is the moment of inertia tensor and is given by
I I xy I xz
xx
I = I yx I yy I yz . . . . (5)
I I zy I zz
zx
1
T = .I . . . . . (6)
2
If the components of the angular velocity and the angular momentum vector L are

= i x + j y + k z ,
. . . (7)
L = iLx + jLy + kLz ,
then we write equation (4) as
Lx I xx I xy I xz x

Ly = I yx I yy I yz y
L I I zy I zz z
z zx
iLx + jLy + kLz = i ( I xx x + I xy y + I xz z ) + j ( I yx x + I yy y + I yz z ) +

+ k ( I zx x + I zy y + I zz z ) .

1
T=
2
( ix + j y + kz ) i ( I xxx + I xy y + I xzz ) + j ( I yxx + I yy y + I yzz ) +
+ k ( I zxx + I zy y + I zz z ) .
This equation becomes

1
T=
2
( I xx x2 + I yy y2 + I zz z2 + 2 I xy x y + 2 I yz y z + 2 I zx z x ) . . . (8)
If the body rotates about z-axis with angular velocity , we have

z = , x = y = 0 . In this case equation (8) becomes
1
T= I . 2 . . . (9)
2
where I is the moment of inertia of the body about z axis.
Worked Examples
Example 11 : Prove the statement that a change in time dt of the components of a

vector r as seen by an observer in the body system of axes will differ from the
corresponding change as seen by an observer in the space system.

Solution: consider two co-ordinate systems S and S , where S is rotating
uniformly with angular velocity , with respect to the frame S with the same origin
O. Let S be the space set of axes and S the body set of axes. Let i, j, k be the unit
vectors associated with the co-ordinate axes of S frame and i, j , k be the unit
z z
vectors associated with x, y, z axes of the
S frame.
k k
P
Consider the position vector r of a particle
r y
j in a rigid body with respect to the body set of axes.
y
i O j It is represented by
x
i r = ix + j y + k z . . . (1)
x
Clearly such a vector appears constant when measured in the body set of
axes. However, to an observer fixed in space set of axes, the components of the
vector will vary in time. Let the components of the vector with respect to the space
set of axes be given by
r = ix + jy + kz . . . . (2)
The time derivatives of r however will be different in the two systems. For
the space (fixed) system S, we have
dr dx dy dz
=i + j +k . . . . (3)
dt fix dt dt dt
Similarly, the time derivative of the vector r defined in S with respect to

the body set of axes is given by
dr dx dy dz
= i + j + k . . . . (4)
dt body dt dt dt
However, as body rotates, the unit vectors of the body set of axes will be seen
changing relative to the observer in the space set of axes. Hence we find the time
derivative of the vector r in S with respect to the fixed co-ordinate system as

dr dx dy dz di dj dk
= i + j + k + x + y + z . . . . (5)
dt fix dt dt dt dt dt dt
The first three terms on the R. H. S. of equation (5) are the time derivative of the
vector in the rotating system, while the remaining three terms arise as a result of
rotation of the system.
Hence we write the equation (5) as
dr dr di dj dk
= + x + y + z . . . . (6)
dt fix dt body dt dt dt
We know the linear velocity of a particle having the position vector r and rotating
with angular is given by
dr
v= =r . . . . (7)
dt
This formula can be applied to the unit vectors as a special case. Thus we write
di dj dk
= i, = j , = k . . . . (8)
dt dt dt
Hence equation (6) reduces to
dr dr
= + x ( i x + j y + k z ) i + y ( i x + j y + k z ) j +
dt fix dt body .
+ z ( i x + j y + k z ) k
dr dr
= + i ( z y y z ) j ( z x x z ) + k ( y x x y ) . . . (9)
dt fix dt body
This is equivalent to
i j k
dr dr
= + x y z .
dt fix dt body
x y z

dr dr
= + r . . . .(10)
dt fix dt body
This is the required relation between the two time derivatives of a vector with respect
to two frames of references.
Note : The formula (10) can also be represented as

dr dr
= +r . . . .(11)
dt space dt rot
Eulers Equations of Motion :

A set of equations governing the rotation of a rigid body referred to its own
axis are known as Eulers equations of motion of a rigid body with one point fixed.
Theorem 6 : Obtain the Eulers equations of motion of a rigid body when one point
of the body remains fixed.
Proof : Consider a rigid of which one point is fixed. Hence translational motion of
the body is absent and the body rotates about an axis passes through the fixed point.
The rotation of the body takes place under the action of torque acting on it. Thus the
equation of the rotational motion of the body in a fixed frame is given by
Torque = rate of change of angular momentum.
dL
N = .
dt fix
However, we know
dr dr
= + r .
dt fix dt body

Therefore the equation of motion of the rigid body becomes
dL dL
N = = + L . . . . (1)
dt fix dt body
where is the angular velocity of the body and L is the angular momentum and is
given by

L = I , . . . (2)

I is the moment of inertia tensor and is constant with respect to the body frame of
reference. We choose the principal axis of the body with respect to which the off
diagonal elements of the moment of inertia tensor vanish and only the diagonal

elements remain in the expression for I . If I1 , I 2 , I 3 are the principal moments of
inertia then we have
I 0 0
1
I =0 I2 0.
0 0 I 3

In this case the expression for angular momentum (2) becomes
L = I1 x i + I 2 y j + I 3 z k ,
Differentiating this with respect to time t in the body frame we get

dL
= I1 x i + I 2 y j + I 3 z k .
dt body
Also we find the value of
i j k
L = x y z
I1 x I 2 y I 3z
L = i ( I 3 I 2 ) y z j ( I 3 I1 ) x z + k ( I 2 I1 ) y x .

Hence the equation of rotational motion of the rigid body becomes
dL
N = + L = i I1 x + ( I 3 I 2 ) yz + j I 2 y + ( I1 I 3 ) x z +
dt body
+ k I 3 z + ( I 2 I1 ) x y
If torque is expressed in term of its components as

N = iN x + jN y + kN z
then on equating the corresponding components on both sides of the above equations
we obtain
N x = I1 x + ( I 3 I 2 ) yz ,
N y = I 2 y + ( I1 I 3 ) x z , . . . (3)
N z = I 3 z + ( I 2 I1 ) x y .
These are the required Eulers equations of motion of the rigid body with one point
fixed.
Note : In the case of torque free motion of a rigid body, equations (3) reduce to
I1 x = ( I 2 I 3 ) y z ,
I 2 y = ( I 3 I1 ) xz , . . . (4)
I 3 z = ( I1 I 2 ) x y .
Worked Examples
Example 12 : If the rigid body with one point fixed rotates about the principal axis
of the body, then show that
(1) kinetic energy of the body and
(2) the magnitude of the angular momentum are constants throughout the motion.
Solution : The kinetic energy of a rigid body with one point fixed is given by
1
T=
2
( I xx x2 + I yy y2 + I zz z2 + 2 I xy x y + 2 I yz y z + 2 I zx z x ) . . . (1)

where I xx , I yy , I zz and I xy , I yz , I zx are the moments of inertia and product of inertia
about the co-ordinate axes respectively. If the body rotates about the principal axis of
the body then the products of inertia tensors are zero and the moments of inertia
tensors are constants. In this case the kinetic energy becomes
1
T=
2
( I1 x2 + I 2 y2 + I 3 z2 ) . . . (2)
where I1 = I xx , I 2 = I yy , I 3 = I zz .
In the absence of torque the Eulers equations of motion of the rigid body are given
by
I1 x + ( I 3 I 2 ) y z = 0,
I 2 y + ( I1 I 3 ) x z = 0, . . . (3)
I 3 z + ( I 2 I1 ) x y = 0.
i) Multiply each equation in (3) by x , y , z respectively and adding we get
I1 x x + I 2 y y + I 3z z = 0 . . . (4)
We write this as
1 d
2 dt
( I1x2 + I 2 y2 + I3z2 ) = 0
dT
= 0,
dt
1
T = ( I1 x2 + I 2 y2 + I 3 z2 ) = const.
2
ii) Now we claim that the magnitude of the angular momentum is constant.
The moment of inertia tensor with respect to the principal axis of the body, is given
by
I 0 0
1
I =0 I2 0,
0 0 I 3


where I1 , I 2, I 3 are constant with respect to the principal axis of the body. Hence the
expression for angular momentum becomes

L = I1 x i + I 2 y j + I 3 z k ,
L2 = L.L = I12 x2 + I 22 y2 + I 32z2 . . . . (5)
Multiply each equation of (3) by I1 x , I 2 y , I 3 z respectively and adding we get
I12x x + I 22 y y + I 32 z z = 0 .
This we write as
1 d 2 2
2 dt
( I1 x + I 22 y2 + I32z2 ) = 0,
d 2

dt
( L ) = 0,
L2 = const.
This shows that the magnitude of the angular momentum of the rigid body is
constant.
Components of angular velocity vector along body set of axes :
Example 13 : Show that the components of angular velocity vector along the body
set of axes are given by
x = sin sin + cos ,
y = sin cos sin ,
z = cos + .
Solution : Let (x, y, z) and ( x, y, z ) be the space (fixed) set of axes and body
(rotating) set of axes respectively. Let a rigid body with one point fixed rotate
instantaneously with angular velocity . We shall obtain the components of
along the body set of axes. If , , are the Eulerian angles, then their time
derivatives ,, represent the angular speeds about the space z-axis, the line of
nodes and the body z-axis respectively. We denote these three angular speeds by

, , which are the three components of the angular velocity . Note that these
three components of are not all either along the space set of axes or the body set of
axes. Let = ( x , y , z ) be the components of with respect to the body set of
axes x, y, z .
If ( ) represents the angular speed about space z-axis then its
components along the body set of axes are found by applying orthogonal
transformations C through an angle about new x -axis and B through an angle
about new z -axis, to come to the body axes, as two orthogonal transformations are
required to come to body axes. If ( ) , ( ) , ( )
x y z are the components of
( ) along body set of axes, then we have
( )
x 0
( ) = BC 0 ,
y

( )
z
cos sin 01 0 0 0

= sin cos 0 0 cos sin 0
0 0 1 0 sin cos

( )
x cos sin cos sin sin 0
( ) = sin cos cos

sin cos 0 ,
y
0 sin cos
( ) z
( )
x = sin sin ,
( )
y
= cos sin , . . . (1)
( )
z = cos .

( )
Similarly, if represents the angular speed along the line of nodes
(new x -axis), and ( ) x , ( ) y , ( ) z are the components of about the( )

body set of axes x, y, z , then to find these components, we apply orthogonal
transformation B-through an angle about new z -axis to come to the body axes;
after rotation has been performed. Thus we have
( )
x
( ) y = B 0 ,
0
( )
z
cos sin 0

= sin cos 0 0
0 0 1 0

( ) x = cos ,
( ) y = sin , . . . (2)
( ) z = 0.
Now ( ) is already parallel to z axis, no transformation is necessary.
Hence, if ( ) x , ( ) y , ( ) z are components of with respect to the body set of
axes, then they are given by
( )
x
= 0,
( )
y
= 0, . . . (3)
( )
z
= .
Thus the components of angular velocity , ( x , y z ) about body set of axes are
defined by

x = ( ) x + ( ) x + ( ) x ,
y = ( ) y + ( ) y + ( ) y ,
z = ( ) z + ( ) z + ( ) z .
Using equations (1), (2) and (3), we readily obtain the components of angular
velocity about body set of axes in the form
x = sin sin + cos ,
y = sin cos sin , . . . (4)
z = cos + .
Example 14 : If a rectangular parallelepiped with its edges 2a, 2a, 2b rotates about
its center of gravity under no force, prove that, its angular velocity about one
principal axis is constant and about the other axis it is periodic.
Solution: It is given that the rigid body rotates under the action of no forces. Hence
the Eulers equation of motion, in the absence of no forces are given by
I1 x = ( I 2 I 3 ) y z ,
I 2 y = ( I 3 I1 ) xz , . . . (1)
I 3 z = ( I1 I 2 ) x y .
where I1 , I 2 , I 3 are called the principal moments of inertia about the center of gravity
of the body. Since the rigid body is parallelepiped with its edges 2a, 2a, 2b. Hence
the moments of inertia about the principal axes OX, OY, and OZ are given by
a2 + b2
I1 = I 2 = m
3 . . . (2)
2
I 3 = ma 2
3
where m is the mass of the parallelepiped. Substituting these values in equation (1)
we get

(a 2
+ b 2 ) x = ( b 2 a 2 ) y z ,
(a 2
+ b 2 ) y = ( a 2 b 2 ) x z , . . . (3)
z = 0 z = n = const.
The last equation in (3) shows that the angular velocity about one principal axis is
constant. Consequently, the other two equations give
(a 2
+ b 2 ) x = ( b 2 a 2 ) n y ,
. . . (4)
(a 2
+ b 2 ) y = ( a 2 b 2 ) nx .
Eliminating y between (4) we obtain

2
a 2 b2
x = n 2 2 x . . . . (5)
a +b
This is a second order differential equation of simple harmonic motion. This shows
that x is periodic.
Unit 3: Caley-Klein Parameters:

Introduction:
We have seen that the Eulerian angles are used to describe an orientation of a
rigid body. However, it is found that these angles are difficult to use in the numerical
computation, because of the large number of trigonometric functions involved.
Various other groups of variables have been used to describe the orientation of a
rigid body. Kliens set of four complex parameters is one of them. He introduced the
set of four parameters bearing his name to facilitate the integration of complicated
gyroscopic problems. These parameters are much better adapted for use on
computers. Furthermore these four parameters are of great theoretical interest in
modern branches of physics. Cayley-Klein parameters are the set of four complex
numbers used to describe the orientation of a rigid body in space.

We begin this unit with some basic definitions and see how these Cayley-
Klein parameters define the orientation of the rigid body.
Some Definitions:
1. Conjugate matrix: The matrix obtained from any given matrix A by
replacing its elements by the corresponding conjugate complex numbers is
called the conjugate of A and it is denoted by A* .
2. Trace of a matrix : Let A be a square matrix of order n. The sum of the
elements of A lying along the principal diagonal is called the trace of A.
3. Transposed conjugate of a matrix : The transpose of the conjugate of a
matrix Q is called transposed conjugate of Q and it is denoted by Q . It is
T *
also called as adjoint of Q. Thus Q = ( Q* ) = ( QT ) .
4. Unitary Matrix : A square matrix Q with complex elements is said to be

unitary if QQ = I = QQ . Unitary condition expect that Q = 1 Q is
invertible and it is given by Q 1 = Q . This also gives that Q 1 = Q = adjQ .

5. Self adjoint : A linear operator which is identical with its adjoint operator is
called self-adjoint. If P is self- adjoint then P = adjP .
6. Hermitian Matrix : A square matrix A = ( aij ) is said to be Hermitian if
aij = a*ji for every i, j . If A = A* A is Hermitian.
7. Similar matrices : Let A and B two square matrices of the same order. Then
A and B are said to be similar if there exists a non-singular matrix P such that
AP = PB. A = PBP 1 .
Property : Under similar transformation the self-adjoint (Hermitian) property of the
matrix and the trace of the matrix are invariant.

Worked Examples
Example 15 : Explain how the eight quantities of Cayley-Klein parameters can be

reduced to only three independent quantities to describe the orientation of a rigid
body?
Solution: Consider a two dimensional complex space with u and v as complex
axis. A general linear transformation in such a space is given by
u = u + v,
. . . (1)
v = u + v ,
where

Q= . . . (2)

is the rotation matrix in 2-dimensional complex plane, and , , , are four
complex parameters are known as Cayley-Klein parameters. There are eight
quantities in four complex parameters. However, we know that the minimum number
of independent quantities needed to specify the orientation of a rigid body is three.
Thus to reduce eight quantities in equation (2) into three independent quantities, the
matrix Q is restricted by imposing an additional condition that it is unitary. The
unitary condition implies that
QQ = I = QQ . . . (3)
where Q is transposed conjugate of Q known as the adjoint matrix. As Q is unitary,
we have Q = 1 = 1 . . . . (4)
Also expanding equation (3) we get
* *
=I
* *
* + * * + * 1 0
* * =
+ * + * 0 1

* + * = 1,
* + * = 1, . . . (5)
* + * = 0.
We notice that the first two equations of (5) are real while the third is complex.
Therefore equation (5) gives 4 conditions. These 4 conditions plus one condition
given in (4) are totally five conditions on eight quantities. Therefore we are left with
only 3 independent quantities, which are used to describe the orientation of the body.
To calculate those three independent quantities dividing equation (4) by we get
1
= . . . . (6)

Now from the last equation in (5) we have
*
= * . . . . (7)

Thus from (6) and (7) we have
* 1
*
=

( * + * ) 1
*
=

= * , = * . . . . (9)
As a result of (9), the matrix Q takes the form

Q= * *
. . . (10)

with the unitary condition
* + * = 1 . . . . (11)
Hence Q involves only 3 independent quantities, which are used to describe the
orientation of a rigid body.

Matrix of transformation in terms of Cayley-Klein Parameters :
Theorem 7 : Obtain the matrix of transformation in Cayley-Klein parameters, which
specify the orientation of a rigid body.
OR
Show that to each unitary matrix Q in the 2-dimensional complex space there
is associated some real orthogonal matrix of transformation in ordinary 3-
dimensional space.
Proof : To find the matrix of transformation in terms of Cayley-Klein parameter, let
P be a matrix operator in a specialized u, v complex co-ordinate system in a
particular form
z x iy
P= . . . (1)
x + iy z
We notice that the matrix P is trace free and is self -adjont. i.e., P = P , and x, y, z
are real quantities taken as co-ordinates of a point in space. Suppose the matrix P is
transformed to the matrix P by means of the unitary matrix Q in the following way
P = QPQ . . . (2)
where

Q= . . . (3)

is the matrix of transformation in 2- dimensional complex space and Q is a complex

*
transposed conjugate of Q. i.e., Q = ( QT ) . Since Q is unitary, from the unitary
property of Q, we have
QQ = I .
This implies that adjoint of Q is same as its inverse. i.e.,
Q 1 = Q = adj ( Q ) . . . . (4)

P = QPQ 1 . . . . (5)
This shows that P is the similarity transformation of P. It is well known that the
self-adjoint and the trace free property of the matrix are invariant under similarity
transformation. As P is self-adjoint and trace-free, therefore P must be like wise
self-adjoint and trace free. Thus P can have the form
z x iy
P =
x + iy z
where x, y, z are to be determined.
Let us denote
x iy = x ,
x + iy = x+ .
Therefore equation (5) with the help of equations (3) and (4) becomes.
z x z x * *
= .
x+ z x+ z * *
We know that , , , are not independent but are related by the equations
* = , * = ,
Therefore we have
z x z x
P = =
x+ z x+
z
z x z x , z + x
P = =
x+ z x+ + z, x+ z
z x z ( + ) x + x+ , 2 x 2 z 2 x+
=
x+ z 2 z 2 x + 2 x+ , x ( + ) z x+
This is the matrix transformation equation in complex 2-plane. Obviously the matrix
on the r. h. s. is hermitean, proving that the hermitean property is invariant under any
unitary similar transformation.
Solving these equations, we obtain
x+ = 2 x+ 2 x + 2 z ,
x = 2 x+ + 2 x 2 z , . . . (6)
z = x+ x + ( + ) z.
In matrix notation, we write these equations as
x+ x+
2
2 2
2 2
x = 2 x . . . . (7)
z + z

Explicitly, we write equations (6) as
x + iy = ( 2 2 ) x + i ( 2 + 2 ) y + 2 z ,
x iy = ( 2 2 ) x i ( 2 + 2 ) y 2 z ,
z = ( ) x + i ( + ) y + ( + ) z.
Solving these equations for x, y, z we ready obtain
1 2 i
2 ( + ) ( 2 2 + 2 + 2 )
2 2 2

x 2
x
i 2
+ + + ) i ( + ) y . . . . (8)
1 2
y = 2 ( + ) (
2 2 2 2 2 2
2
z z

i ( + ) +

Thus we have
1 2 i
2 ( + ) , ( 2 2 + 2 + 2 ) ,
2 2 2

2

1 2
A = ( 2 2 2 + 2 ) , + + + ) , i ( + ) .
i
2 2
( 2 2 2

. . . (9)

, i ( + ) , +

This is the required matrix of transformation in terms of Cayley-Klein parameters.
This matrix specifies the orientation of a rigid body. Hence the Cayley-Klein
parameters specify the orientation of a rigid body.

Relation between the Cayley-Klein Parameters and Eulerian Angles :
Theorem 8 : Establish the relation between the Eulerian angles , , and the
Cayley-Klein parameters , , , .
OR
Obtain Q matrices Q , Q , Q in complex 2-plane corresponding to the separate
successive rotations through an angle , , in 3 3 real space. Hence obtain the

orthogonal matrix of complete rotation.
Proof : The Eulerian angles , , are the three successive angles of rotations about
the specified axes, such that the space set of co-ordinates (x, y, z) coincide with the
body set of co-ordinates ( x, y, z ) . These angles are used to describe the orientation
of a rigid body. Similarly, Cayley-Klein parameters , , , are also used to

describe the orientation of rigid body. Now to find the relation between , , and
, , , , we first construct Q matrices say Q , Q , Q corresponding to the
separate successive rotations , , and then combine them to form the complete
matrix Q = Q Q Q of rotation.
First Rotation :

Let Q = be the matrix in 2-complex plane corresponding to the first

rotation through an angle in 3 3 real space. This rotation through an angle is
performed about z-axis. Hence the transformation equations are given by
x = x cos + y sin ,
y = x sin + y cos ,
z = z
We write these equations as

x+ = x + iy = xe i + iye i
x+ = e i x+ ,
. . . (1)
x = ei x ,
z = z.
But we know that
x+ = 2 x+ 2 x + 2 z ,
x = 2 x+ + 2 x 2 z , . . . (2)
z = x+ x + ( + ) z.
Now comparing equations (1) and (2) we get

2 = ei , 2 = e i , = 0 = .
Therefore the Q matrix corresponding to the rotation through an angle becomes
i2
e 0
Q = i
. . . . (3)
0
e 2

Second Rotation :

Let Q = be the matrix in 2-complex plane corresponding to the

second rotation through an angle in 3 3 real space. This rotation through an angle
is performed about new x-axis. Hence the transformation equations are given by
x = x,
y = y cos + z sin , . . . (4)
z = y sin + z cos .
From equations (4) we obtain

x+ = x + iy = x cos 2 + sin 2 + iy cos + iz sin
2 2

x = x iy = x cos 2 + sin 2 iy cos iz sin
2 2
z = y sin + z cos .

x+ = x cos 2 + sin 2 + iy cos 2 sin 2 + 2iz sin cos .
2 2 2 2 2 2

x+ = x+ cos 2 + x sin 2 + 2iz sin cos .
2 2 2 2
Similarly, we write

x = x+ sin 2 + x cos 2 2iz sin cos ,
2 2 2 2
. . . (5)

z = ix+ sin cos ix sin cos + z cos 2 sin 2 .
2 2 2 2 2 2
Comparing equations (5) with (2) we get

2 = 2 = cos 2 , 2 = 2 = sin 2 .

2
2
Hence Q matrix becomes

cos 2 , i sin 2

Q = . . . . (6)

i sin , cos
2 2

Third Rotation: Let Q = be the matrix in 2-complex plane corresponding

to the third rotation through an angle in 3 3 real space. This rotation through an
angle is performed about new z-axis. This rotation is affected by the
transformation equations
x = x cos + y sin ,
y = x sin + y cos ,
z = z

We write these equations in the form
x+ = e i x+ ,
x = ei x , . . . (7)
z = z.
Hence comparing equation (7) with (2) we obtain the matrix Q corresponding to
the third rotation through an angle about new z- axis as
i2
e , 0
Q = i . . . . (8)
0, e 2

Hence the orthogonal matrix for the complete transformation from space set of axes
to the body set of axes is obtained by taking the product of separate Q matrices for
each of the three successive rotations , , . Thus we obtain
Q = Q Q Q ,

i2 cos , i sin i2
e , 0 2 2 e , 0
Q= ,
i sin , cos 0, e 2
i i
0, e 2

2 2
2i ( + )
i
( )
sin
2
e cos , ie
2 2
Q= = . . . . (9)
2i ( ) i
( + )
ie sin , e 2
cos
2 2
i
( + )
= e2 cos ,
2
i
( )
= ie 2 sin ,
2

i
( )
= ie 2 sin ,
2
i
. . . (10)
( + )
=e 2
cos .
2
These are the required relations between the Eulerian angles and the Cayley-Klein
parameters.
Note: Note from equations (3), (6) and (8) that the trace of any Q matrix through an

angle say about some axis is 2 cos .
2
Worked Examples
Example 16 : (Aliter) With usual notations show that

+
cos = cos cos .
2 2 2
Solution: This example is solved in Example (9). However, we will attempt this
problem by using the relation between Eulerian angles and Cayley-Klein parameters.

We know the Q matrix in terms of Eulerian angles is given by

2i ( + )
i
( )
sin
2
e cos , ie
2 2
Q= = . . . (1)
2i ( ) i
( + )
ie sin , e 2 cos
2 2
where , , , are Cayley-Klein parameters such that * = , * = .

This is the matrix of rotation, which describes the orientation of the rigid
body with one point fixed. Let this matrix be equivalent to the matrix B obtained by
rotating the co-ordinate axes through an angle about any axis with the same
origin. Then the matrix B is given by

cos sin 0 1 0 0

B = sin cos 0 or B = 0 cos sin . . . . (2)
0 0 1 0 sin cos

Then the Q matrix in complex 2-dimensional plane corresponding to the matrix B is
similarly obtain in the form

i2 cos , i sin
0 2 2
Q =
e
i
or Q = . . . . (3)

0 e 2
i sin , cos
2 2
From equation (1) we have
i
( + )
= e0 + ie3 = e 2 cos ,
2
i
( + )
= e0 ie3 = e 2
cos ,
2
i
. . . (4)
( )
= e2 + ie1 = ie 2
sin ,
2
i
( )
= e2 + ie1 = ie 2
sin .
2
Solving these equations, we find
+ +

e0 = = cos cos ,
2 2 2
i
e1 = ( + ) = sin cos ,
2 2 2
. . . (5)

e2 = = sin sin ,
2 2 2
+
e3 = = cos sin .
2 2 2
We see that
+
trace of Q = + = 2 cos cos . . . . (6)
2 2

Similarly, from equation (3) we have

trace of Q = 2 cos . . . (7)
2
Since trace is invariant, therefore, from (6) and (7) we have
+
cos = cos cos .
2 2 2
Example 17 : Find a real matrix of orthogonal transformation in the 3-dimensional

space corresponding to the unitary matrix

cos 2 , i sin 2

Q=

i sin , cos
2 2
in two dimensional complex plane.
Solution: To find the matrix of orthogonal transformation in 3-dimensional space
corresponding to the matrix

cos 2 , i sin 2

Q= . . . (1)

i sin , cos
2 2
in 2-dimensional complex plane, let P be a required matrix operator in a specialized
u, v complex co-ordinate system in a particular form
z x iy
P= . . . (2)
x + iy z
We notice that the matrix P is trace free and is self -adjont. i.e., P = P , and x, y, z
are real quantities taken as co-ordinates of a point in space. Suppose the matrix P is
transformed to the matrix P by means of the unitary matrix Q in the following way
P = QPQ . . . (3)

*
where Q is a complex transposed conjugate of Q. i.e., Q = ( QT ) called adjoint of
Q. Since Q is unitary, from the unitary property of Q, we have QQ = I .

This implies that adjoint of Q is same as its inverse. i.e.,
Q 1 = Q = adj ( Q ) . . . . (4)

P = QPQ 1 . . . . (5)
This is the similarity transformation of P. Under similar transformation, we know
that the trace-free and self-adjoint property of the matrix are invariant. Since P is
self-adjoint and trace-free, therefore P must be like wise self-adjoint and trace free.
Thus P can have the form
z x iy
P = . . . (6)
x + iy z
where x, y, z are to be determined.
Let us define
x iy = x ,
. . . (7)
x + iy = x+ .
Therefore equation (5) with the help of equations (1), (2), (6) and (7) becomes.

cos i sin cos i sin
z x 2 2 z x 2 2
= .
x+ z x+ z
i sin cos i sin cos
2 2 2 2

Therefore we have
z x
=
x+ z
2 2
zcos sin +ix+cos sin ixcos sin , x sin2 +x cos2 2izcos sin
2 2 2 2 2 2 + 2 2 2 2
= .

x cos2 +x sin2 +2izcos sin , zsin2 cos2 +ix cos sin ix cos sin
+ 2 2 2 2 2 2 2 2 + 2 2
On equating the corresponding elements of the matrix we get

x+ = x + iy = 2iz cos sin + ( x + iy ) cos 2 + ( x iy ) sin 2 ,
2 2 2 2

x = x iy = 2iz cos sin + ( x iy ) cos 2 + ( x + iy ) sin 2 ,
2 2 2 2

z = z cos 2 sin 2 i ( x iy ) cos sin + i ( x + iy ) cos sin .
2 2 2 2 2 2
x + iy = x + iy cos + iz sin ,
x iy = x iy cos iz sin ,
z = z cos y sin .
x = x
y = y cos + z sin ,
z = y sin + z cos .
x 1 0 0 x

i.e., y = 0 cos sin y
z 0 sin cos
z
This shows that corresponding to the matrix

cos 2 i sin 2

Q=

i sin cos
2 2

in 2-dimensional complex plane, there exits 3 3 real matrix
1 0 0

A = 0 cos sin
0 sin cos

in 3- dimensional real space.
Exercise:
1. Show that the components of angular velocity vector along the space set of
axes are given by
x = cos + sin sin ,
y = sin sin cos ,
z = cos + .
2. Find a real matrix of orthogonal transformation in the 3-dimensional space
corresponding to the unitary matrix
i2
e 0
Q= i
0 e 2

in 2- dimensional complex plane.
3. Find a real matrix of orthogonal transformation in the 3-dimensional space
corresponding to the unitary matrix
i2
e 0
Q= i
0 e 2

in 2- dimensional complex plane.

Langrangian Hamilton

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Langrangian Hamilton

Uploaded by

Copyright:

Available Formats

CHAPTER - I

. the bullet at that instant and also

Classical Mechanics Page No. 1

Classical Mechanics Page No. 2

the position vector, then velocity of the particle is v = r = xi

Classical Mechanics Page No. 3

L = r p . Angular momentum is perpendicular to both the position vector and the

Classical Mechanics Page No. 4

Classical Mechanics Page No. 5

Equation of motion of a system of particles:

vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. Any particle of this

ii) Internal forces Fji( int ) .

Classical Mechanics Page No. 6

presence of remaining n-1 particles. However, there is no self interacting force,

Fji(int ) = Fij( int ) .

Thus equation of motion (2) of the system becomes

vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. The angular momentum of

the i th particle of the system about the origin is given by

Classical Mechanics Page No. 7

However, the term r F(

Interchanging i and j on the r. h. s. of equation (5) we get

Classical Mechanics Page No. 8

Adding equations (5) and (6) we get

Classical Mechanics Page No. 9

vectors ri , i = 1, 2,..., n relative to an arbitrary fixed origin. The angular momentum of

the i th particle of the system about the origin is given by

where vi is the velocity of the i th particle with respect to O,

vi - velocity of the i th particle with respect to centre of mass,

v - velocity of the centre of mass with respect to O.

Classical Mechanics Page No. 10

Consequently we have from above equation

ordinates of the i th particle of the system, i = 1, 2,..., n .

Classical Mechanics Page No. 11

then constraints are said to be holonomic otherwise non-holonomic constraints. A

Hence such constraints are also called non-holonomic constraints. Obviously,

Classical Mechanics Page No. 12

Classical Mechanics Page No. 13

Definition : The least possible number of independent co-ordinates required to

Classical Mechanics Page No. 14

q j , j = 1, 2,..., n are the generalized co-ordinates, then there exists a relation

called the transformation relation.

the work done by F is given by

Classical Mechanics Page No. 15

Example 3 : Show that the gravitational force is conservative.

W = mg If F = Xi + Yj and r = xi + yj , where X and Y

where a and b are the ordinates at points P and Q respectively.

However, by Stokes theorem, we have

Classical Mechanics Page No. 16

For conservative force we have F = 0.

Classical Mechanics Page No. 17

Virtual displacement ri is assumed to take place only in the co-ordinates and at

Classical Mechanics Page No. 18

Principle of Virtual Work :

pi is the linear momentum of the i th particle. This can be written as Fi p i = 0 .

implying a system of particles is in equilibrium. This equation states that the

Classical Mechanics Page No. 19

Example 5: A particle is constrained to move in a circle in a vertical plane xy. Apply

point 0.The constraint on the motion of the particle is that

( Fx mx) x + ( Fy my) y = 0 . . . . (3)

Classical Mechanics Page No. 20

and Fy = mg . Therefore equation (3) becomes

On using (2) we have

For x 0, and m 0 we have

Example 6: Use DAlemberts principle to determine the equation of motion of a