Professional Documents
Culture Documents
Learning Outcomes:
After reading this lesson, students will be able to:
understand the concept of Duality in linear programming,
write a dual problem from a given primal problem,
read solution of the dual problem from the solution of the primal
problem,
understand the concept of shadow prices,
perform the economic interpretation of Dual, and
understand post optimality analysis in LPP.
1. Introduction:
For every Linear Programming Problem there exists a Dual that expresses the problem
with a resource orientation. The original problem is called the Primal Problem. The
solution of a Linear Programming Problem can be read from the final simplex tableau of
either Primal or its Dual for all the variables.
The importance of duality in linear programming is twofold. Firstly, it is often possible to
solve the dual linear problem in place of the original linear problem (Primal), to reap
advantage of computational efficiency. Secondly, it helps in understanding the shadow
price interpretation of the optimal simplex solution.
At times it is easier to solve the dual problem instead of the Primal Problem,
because of easier computation.
For instance, in case we have all constraints in Primal requiring Surplus as
well as Artificial variables to be introduced for each of the constraint, it will
be easier to solve the Dual and read the solution of Primal from the optimal
final tableau of the Dual. The Optimal value of the Objective Function is the
same for both the Primal and its associated Dual.
Dual of a Dual is the Primal itself.
Illustration 4:
Write the Dual of the following LPP:
Maximize Z = 35x1+29x2+11x3
x1, x2, x3 0
Solution:
Dual Problem:
Let 1 , 2 , 3 be the dual variables
Minimize C = 100y1+150y2
10y1+ 20y2 35
25y1+ 5y2 29
40y1+ 30y2 11
y1, y2 0
Primal Dual
Maximization Minimization
Objective function
Minimization Maximization
Number of
n m
decision variables
Coefficients of
objective function
Constraints
Number of constraints m n
RHS Coefficients s s
Technological
A matrix At matrix
coefficients
Optimal Value of Zj Value in Zj Value in
Objective Function Quantity Column Quantity Column
Optimal Solution of
Decision Variables of Quantity or column
(under slack variables)
Primal
Optimal Solution of
Decision Variables of Quantity or column
(under slack variables)
Dual
Illustration 5:
Write Dual of the following LPP:
Minimize Z = 121 + 152 + 173
Subject to constraints:
51 92 + 3 18
111 23 = 15
22 + 193 10
1 , 3 0 2 :
Solution:
Dual Problem:
Let 1 , 2 , 3 be the dual variables
Maximize = 181 + 152 103
Subject to constraints:
51 + 112 12
91 23 = 15
1 22 193 17
1 , 3 0 2 :
Illustration 6:
Find the dual to the following LPP:
Maximize Z = 4 x1 2 x2 3x3
Illustration 7:
Write a dual for the following linear programming problem:
Maximize Z = 4 x1 5x2 7 x3
Subject to the constraints
x1 x2 x3 10
4 x1 x2 2 x3 15
x1 4 x2 7 x3 35
x1 , x2 0 , x3 is unrestricted.
University of Delhi, B.Com. Hons.2011
Solution:
Dual Problem:
Let 1 , 2 , 3 be the dual variables.
Minimize = 101 152 353
Subject to constraints
1 42 + 3 4
1 + 2 43 5
1 22 73 = 7
1 , 2 , 3 0
Illustration 8:
Write a dual to the following LPP:
Maximize Z = 20 x1 15x2 18x3 10 x4
Subject to the constraints
4 x1 3x2 10 x3 4 x4 60
x1 x2 x3 27
x2 4 x3 7 x4 35
x1 , x2 , x3 0 and 4 :
University of Delhi, B.Com Hons.2006 (CC)
Solution:
Dual Problem: Let 1 , 2 , 3 be the dual variables.
Minimize = 601 + 272 353
Subject to constraints
41 + 2 20
31 + 2 + 3 15
101 + 2 43 18
41 73 = 10
1 , 3 0 and 2 :
Source: https://www.youtube.com/watch?v=rO-ufnCNqXM
4. Primal-Dual Solution:
The optimum solution for a dual can be read from the Final Simplex tableau of the Primal
problem and vice versa. At times it is easier to solve the dual and read solution for the
Primal. For instance, if we have a two variable minimization LPP, with four constraints,
then for solving the problem we will have to add eight constraints as surplus and
artificial variables. However, for the maximization with all constraints we will have to
add slack variables equal in number to the constraints. Therefore, it will be
computationally efficient to write the Dual of the problem and solve it to read solution of
the Primal.
Optimal Solution of the Dual
The optimal value of objective function is same for both the Primal and the Dual
solutions.
Solution of the decision variables of the Dual
The solution for the decision variables in the Dual can be read from the Final Simplex
Tableau of the Primal Problem as given below:
(i) Variable in the Dual corresponding to the constraint in the Primal:
absolute value of the entry in Cj-Zj row under the corresponding surplus variable.
(ii) Variable in the Dual corresponding to the = constraint in the Primal:
corresponding Zj value for Artificial variable.
(iii) Variable in the Dual corresponding to the <= constraint in the Primal:
absolute value of entry in Cj-Zj row under the corresponding variable.
The dual variables represent various constraints in the primal problem, therefore their
values in the final simplex solution reflects the shadow prices of the corresponding
resources.
Illustration 9:
A manufacturer produces two products A and B and earns a profit of Rs 8 on each unit of
A and Rs 7 on each unit of B. The required machines M1, M2 and M3 are available for 30
hours, 60 hours and 50 hours respectively. Product A needs 1 hour on M1, 2 hours on
M2 and 1 hour on M3. Product B needs 2 hours on M1, 3 hours on M2 and 1 hour on M3.
You are required to:
Formulate the above as Linear Programming Problem and write its associated
dual.
Solve both the primal and dual problems.
Solution:
Formulation of Primal Problem:
Let the number of units of product A and product B manufactured are 1 2
respectively.
Maximize = 81 + 72 (Total profit)
Subject to constraints:
1 + 22 30 (Availability of Machine M1)
21 + 32 60 (Availability of Machine M2)
1 + 2 50 (Availability of Machine M3)
1 , 2 0
Simplex Solution:
Introducing the slack variables 1 , 2 , 3 :
Maximize = 81 + 72 + 01 + 02 + 03
Subject to constraints:
1 + 22 + 1 = 30
21 + 32 + 2 = 60
1 + 2 + 3 = 50
1 , 2 , 1 , 2 , 3 0
Augmented LPP:
Maximize = 21 + 52 + 01 + 02 + 03
Subject to constraints
1 + 22 + 1 + 02 + 03 = 30
21 + 32 + 01 + 2 + 03 = 60
1 + 2 + 01 + 02 + 3 = 50
1 , 2 , 1 , 2 , 3 0
Initial Basic Feasible Solution:
An initial basic feasible solution is obtained by putting the decision variables equal to
zero and calculating values of the slack variables.
Putting 1 = 0 2 = 0
1 = 30, 2 = 60, 3 = 50 and Z=0
8 7 0 0 0
Basic Solution Ratio=
Variable
0 1 30 1 2 1 0 0 30
0 2 60 2 3 0 1 0 30
0 3 50 1 1 0 0 1 50
0 0 0 0 0 0
8 7 0 0 0
Incoming Variable: x1 is the incoming variable because it has the highest C j - Z j value
in the Net Evaluation Row, i.e. 8. The column containing it is the key column which is
used to calculate the replacement ratios in the ratio column.
Outgoing Variable: s1 is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 30. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 1 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-II containing the improved
solution is obtained as under:
Simplex Tableau-II
8 7 0 0 0
8 1 30 1 2 1 0 0
0 2 0 0 -1 -2 1 0
0 3 20 0 -1 -1 0 1
240 8 16 8 0 0
0 -9 -8 0 0
The Net Evaluation Row indicates that the C j - Z j values are non-positive for all the
variables, thus the solution is optimal.
Optimal solution:
x1 =30 and x2 = 0
Maximum value of the objective function, Z=240.
Thus, 30 units of Product A and 0 units of Product B will be manufactured yielding
the maximum total profit of Rs 240.
Dual Problem:
Let 1 , 2 and 3 is the worth (Rs per hour) of the machines M1, M2, M3 used for
producing product A and product.
Minimize = 301 + 602 + 503
Subject to constraints:
1 + 22 + 3 8
21 + 32 + 3 7
1 , 2 , 3 0
Simplex Solution:
Introducing the surplus variables, viz., 1 , 2 and artificial variables viz., 1 , 2 :
Minimize = 301 + 602 + 503 + 1 + 2
Subject to constraints:
1 + 22 + 3 1 + 1 = 8
21 + 32 + 3 2 + 2 = 7
1 , 2 , 3 , 1 , 1 , 2 , 2 0
Augmented LPP:
Minimize = 301 + 602 + 503 + 01 + 02 + 1 + 2
Subject to constraints:
1 + 22 + 3 1 + 02 + 1 + 02 = 8
21 + 32 + 3 + 01 2 + 01 + 2 = 7
1 , 2 , 3 , 1 , 1 , 2 , 2 0
Initial Basic Feasible Solution:
An initial basic feasible solution is obtained by putting the decision variables and the
surplus variables equal to zero and calculating values of the slack and artificial variables.
Putting y1 = 0, 2 = 0, 3 = 0, 1 = 0 and 2 = 0
1 = 8, 2 = 7 and Z=15M
Accordingly, the initial tableau appears as under:
Simplex Tableau-I
30 60 50 0 0 M M
Basic Solution Ratio=
Variable
M 1 8 1 2 1 -1 0 1 0 8/2=4
M 2 7 2 3 1 0 -1 0 1 7/3
15M 3M 5M 2M M M M M
Incoming Variable: is the incoming variable because it has the least C j - Z j value in
the Net Evaluation Row, i.e.5M+60. The column containing it is the key column which
is used to calculate the replacement ratios in the ratio column.
30 60 50 0 0 M M
M 10/3 -1/3 0 1/3 -1 2/3 1 -2/3 =5
7 3
60 7/3 2/3 1 1/3 0 -1/3 0 1/3 =-7
3 1
2 2M
+ 3 +40 60 3
+ 20 M 3
20 M + 20
3
5
10 0
3
+ 30 M
+ 0 3
20
3
Incoming Variable: is the incoming variable because it has the least C j - Z j value in
the Net Evaluation Row, i.e. + . The column containing it is the Key Column
which is used to calculate the replacement ratios in the ratio column.
Outgoing Variable: A1 is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 5. The row containing it is the Key Row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 2/3 that will be used to perform the Elementary Row
Operation to obtain the improved solution. The simplex table-III containing the
improved solution is obtained as under:
Simplex Tableau-III
30 60 50 0 0 M M
1 1 3 3
0 5 0 1 1 ve
2 2 2 2
1 1 1
60 4 1 0 0 8
2 2 2
240 30 60 30 30 0 30 0
0 0 20 30 0 M-30
The Net Evaluation Row indicates that the C j - Z j values are non-negative for all the
variables, thus the solution cannot be minimized further and therefore it is optimal.
Optimal solution:
= , = 4 and = 0
Minimum value of the objective function, Z*=240.
The non-basic variable y1 has zero value in the row which indicates that the
solution is multiple optimal.
30 60 50 0 0 M M
Basic Solution
Variable
0 9 0 1 1 2 1 2 1
30 8 1 2 1 1 0 1 0
30 60 30 30 0 30 0
0 0 20 30 0 M-30
The Net Evaluation Row indicates that the C j - Z j values are non-negative for all the
variables, thus the solution is optimal.
Alternate Optimal Solution:
= , = 0 and = 0
Minimum value of the objective function, Z*=240.
Primal Dual
Basic Basic
Variable Variable
1 30 0 8 0
2 0 -9 0 0
0 -8 0 20
0 0 1 0 30
20 0 2 9 0
Z 240 Z* 240
Illustration 10:
A diet is to contain at least 20 ounces of protein and 15 ounces of carbohydrates. There
are three foods: A, B and C available in the market, costing Rs 2, Rs 1 and Rs 3 per unit
respectively. Each unit of A contains 2 ounces of protein and 4 ounces of carbohydrates.
Each unit of B contains 3 ounces of protein and 2 ounces of carbohydrates. Each unit of
C contains 4 ounces of protein and 2 ounces of carbohydrates.
(i) Formulate LPP so as to minimize the cost of diet.
(ii) Find its dual.
(iii) Solve the dual by simplex method and from the dual solution read the solution to
the primal problem. University of Delhi, B. Com. Hons.2007
Solution:
(i) Formulation of LPP:
Let the units of food A, B and C be 1 , 2 and 3 respectively.
The LPP can be expressed as:
Minimize = 21 + 2 + 33 (Total cost of diet)
Subject to the constraints:
21 + 32 + 43 20 (Protein constraint)
41 + 22 + 23 15 (Carbohydrates constraint)
1 , 2 , 3 0
(ii) Dual Problem:
Let 1 2 be the shadow price of Protein (per ounce) and Carbohydrates (per ounce)
respectively.
The Dual can be written as:
Maximize = 201 + 152
Subject to the constraints:
21 + 42 2 (Food A constraint)
31 + 22 1 (Food B constraint)
41 + 22 3 (Food C constraint)
1 , 2 0
20 15 0 0 0
Basic Solution Ratio=
Variable
2
0 1 2 2 4 1 0 0 =1
2
0 2 1 3 2 0 1 0
3
0 3 3 4 2 0 0 1
4
0 0 0 0 0 0
20 15 0 0 0
Incoming Variable: 1 is the incoming variable because it has the highest C j - Z j value
in the Net Evaluation Row, i.e. 20. The column containing it is the key column which is
used to calculate the replacement ratios in the ratio column.
Outgoing Variable: 2 is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 1/3. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 3 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-II containing the improved
20 15 0 0 0
Basic Solution Ratio=
Variable
/
0 1 4/3 0 8/3 1 2/3 0 =
/
1/3 1
20 1 1/3 1 2/3 0 1/3 0 =
2/3 2
5/3
0 3 5/3 0 2/3 0 4/3 1 =
2/3
20/3 20 40/3 0 20/3 0
0 5/3 0 20/3 0
Incoming Variable: is the incoming variable because it has the highest positive C j -
Z j value in the Net Evaluation Row, i.e. 5/3. The column containing it is the key
column which is used to calculate the replacement ratios in the ratio column.
Outgoing Variable: is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 1/2. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 8/3 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-III containing the
improved solution is obtained as under:
Simplex Tableau-III
20 15 0 0 0
Basic Solution Ratio=
Variable
20 1 0 1 0 -1/4 1/2 0
0 3 2 0 0 0 1 1
0 0 -5/8 25/4 0
The Net Evaluation Row indicates that the C j - Z j values are non-positive for all the
variables, thus the solution is optimal.
Optimal solution: 1 =0 and 2 = 1/2
Maximum value of the objective function, Z=15/2.
Reading the Solution of the Primal under the slack variables in the C j - Z j row:
A 1 3 2 11 15
B 3 4 1 12 20
C 2 2 2 10 16
Hours Available 160 120 180
What should the product mix be? Write the dual of the given problem.
Give its economic interpretation and use it for checking the optimal solution of the given
problem. University of Delhi, B.Com. Hons. 2011
Solution:
Formulation of Primal Problem:
Let the units of products A, B and C manufactured are 1 , 2 and 3 be respectively.
Revenue can be calculated as:
Revenue
Product Selling Price (Rs) Cost (Rs)
(SP-Cost) (Rs)
A 15 11 4
B 20 12 8
C 16 10 6
4 8 6 0 0 0
160
0 1 160 1 3 2 1 0 0 = 53.3
3
0 2 120 3 4 2 0 1 0 =
180
0 3 180 2 1 2 0 0 1 = 180
1
0 0 0 0 0 0 0
4 8 6 0 0 0
Incoming Variable: 2 is the incoming variable because it has the highest C j - Z j value
in the Net Evaluation Row, i.e. 8. The column containing it is the key column which is
used to calculate the replacement ratios in the ratio column.
Outgoing Variable: 2 is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 30. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 4 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-II containing the improved
solution is obtained as under:
Simplex Tableau-II
4 8 6 0 0 0
Basic Solution Ratio=
Variable
5 1 3 70
0 1 70 0 1 0 = 140
4 2 4 1/2
3 1
8 2 30 1 0 0 =
4 4 /
5 3 1 150
0 3 150 0 0 1 = 100
4 2 4 3/2
240 6 8 4 0 2 0
-2 0 2 0 -2 0
4 8 6 0 0 0
Basic Solution
Variable
0 1 40 2 -1 0 1 1 0
3 1
6 3 60 2 1 0 0
2 2
0 3 60 -1 -3 0 0 1 1
360 9 12 6 0 3 0
-5 -4 0 0 -3 0
The Net Evaluation Row indicates that the values are non-positive for all the
variables, thus the solution is optimal.
Optimal solution:
=0, = and =60
Maximum value of the objective function, Z=360.
Economic interpretation of dual:
The dual variables 1 , 2 and 3 represent the shadow prices or the worth of the
constraints corresponding to Centre I (per Hour), Centre II (per Hour) and Centre III
(per Hour), used to manufacture the three products respectively. Z will be the minimum
cost involved in the production.
The shadow prices are Rs 0 per hr., Rs3 per hr. and Rs 0 per hr. for Centre I, II and III
respectively. If the production capacity is to be expanded then the resource having
highest shadow price is given priority. Here Centre II will be given priority.
Centre I and Centre II have unutilized capacity of 40 hours and 60 hours respectively.
Optimal Solution of the Dual:
1 = 0, 2 = 3 and 3 = 0 and Z=360.
Screws(kg) 2 4 3 396kg
Solution:
(i) (a) Formulation of Primal LPP:
Let the number of Row Boat, Canoe and Kayak manufactured are 1 , 2 and
3 respectively.
The LPP can be expressed as under:
Maximize = 40001 + 20002 + 50003 (Total Revenue)
Subject to the constraints:
121 + 72 + 93 1260 (Labour constraint)
221 + 182 + 163 19008 (Wood constraint)
21 + 42 + 33 396 (Screws constraint)
1 , 2 , 3 0
(b) Dual Problem:
Let 1 , 2 and 3 be the shadow price of Labour (per Hour), Wood (per Board feet)
and Screws (per kg.) used to manufacture the boats respectively. The Dual can be
written as:
Minimize = 12601 + 190082 + 3963 (Total shadow price of the resources)
Subject to the constraints:
121 + 222 + 23 4000 (Row Boat constraint)
71 + 182 + 43 2000 (Canoe constraint)
91 + 162 + 33 4000 (Kayak constraint)
1 , 2 , 3 0
(ii) Simplex Solution of the Primal:
Introducing the slack variables 1 , 2 , 3 the LPP can be expressed as:
Maximize = 40001 + 20002 + 50003 + 01 + 02 + 03
Subject to constraints:
121 + 72 + 93 + 1 = 1260
221 + 182 + 163 + 2 = 19008
21 + 42 + 33 + 3 = 396
1 , 2 , 3 , 1 , 2 , 3 0
Augmented LPP:
Maximize = 40001 + 20002 + 50003 + 01 + 02 + 03
Subject to constraints:
121 + 72 + 93 + 1 + 02 + 03 = 1260
221 + 182 + 163 + 01 + 2 + 03 = 19008
21 + 42 + 33 + 01 + 02 + 3 = 396
1 , 2 , 3 , 1 , 2 , 3 0
1260
0 1 1260 12 7 9 1 0 0 = 140
9
19008
0 2 19008 22 18 16 0 1 0 = 1188
16
0 3 396 2 4 3 0 0 1 =
0 0 0 0 0 0 0
4000 2000 5000 0 0 0
Incoming Variable: 3 is the incoming variable because it has the highest value
in the Net Evaluation Row, i.e. 5000. The column containing it is the key column which
is used to calculate the replacement ratios in the ratio column.
Outgoing Variable: 3 is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 132. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 3 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-II containing the improved
solution is obtained as under:
Simplex Tableau-II
0 1 72 6 -5 0 1 0 -3 =
34 10 16896 25344
0 2 16896 0 0 1 -16/3 =
3 3 34/3 17
2 4 132
5000 3 132 1 0 0 1/3 = 198
3 3 2/3
14000 5000
0 0 0
3 3
Incoming Variable: is the incoming variable because it has the highest positive C j -
Z j value in the Net Evaluation Row, i.e. 2000/3. The column containing it is the key
column which is used to calculate the replacement ratios in the ratio column.
Outgoing Variable: is the outgoing basic variable corresponding to the minimum
positive ratio in the ratio column, i.e. 12. The row containing it is the key row.
Pivot Value: The highlighted number at the intersection of the Key Column and the Key
Row is the Pivot Value, i.e. 6 that will be used to perform the elementary row
operation to obtain the improved solution. The simplex table-III containing the
improved solution is obtained as under:
Simplex Tableau-III
Basic Solution
Variable
5
1 12
6
55
2 16760
9
17
3 124
9
55000
668000
9
37000
9
Illustration 12:
Given the simplex tableau for a maximization problem of linear programming:
Product
Quantity
Mix
4 2 1 1 1 0 6
0 2 1 0 -1 1 2
i) Test whether the above table gives optimum solution. If not, improve it and find the
optimum solution.
ii) If s1 and s2 are slack variables at Machine-I and Machine-II, what is the maximum
price you will like to pay for one hour of each machine?
University of Delhi, B.Com. Hons. 2006
Solution:
Simplex Tableau
3 4 0 0
Basic Solution
Ratio=
Variable
4 2 6 1 1 1 0
0 2 2 1 0 1 1
24 4 4 4 0
1 0 4 0
(i) The given solution is optimum solution with 1 =0 and 2 = 6 because the value
of is non-positive for all the variables. Accordingly the objective function
stands maximized at Z=24.
(ii) The maximum price the producer will like to pay for one hour of each machine
is given by the shadow prices. The shadow prices of the Machine I and
Machine II, is the absolute values of and variables in row i.e. Rs 4
per hour of Machine I and Rs 0 for Machine II (which has 2 hours unutilized).
Illustration 13:
Given below are the objective function, the constraints and the final simplex tableau for
a product mix LPP:
Maximize Z = 15x1 + 12x2 + 11x3 + 0S1 + 0S2 + 0S3
Subject to constraints:
3x1 + 4x2 + 5x3 63 (hours, Department A)
4x1 + x2 + 4x3 64 (hours, Department B)
x1 + 4x2 + 2x3 40 (hours, Department C)
x1, x2, x3 0
Final Simplex Tableau
Cj 15 12 11 0 0 0
(i) Write the optimal product mix and the profit contribution shown by the above
solution.
(ii) Is this solution feasible? Why? Give brief reason(s).
(iii) Does the problem have any alternative optimal solution? If so, show one such
solution.
(iv) Indicate the shadow prices of three departments.
(v) If the company wishes to expand the production capacity, which of the three
departments should be given priority.
(vi) If the company produces thirteen units of x3 how many units of x1 and x2 shall
have to be reduced, if any?
(vii) If a customer is prepared to pay higher prices for product x3 how much should the
price be increased so that the company's profit remains unchanged?
(viii) Indicate whether the solution given in the table is degenerate. If yes, which
variable is degenerated?
Solution:
60 193
(i) The optimal product mix is 1 = , 2 = 3 = 0 . The profit
13 12
3615
contribution is Rs. .
13
(ii) This solution feasible because there is no artificial variable in the optimal solution.
(iii) No, the problem does not have any alternative optimal solution because the non-
basic variables viz., x3, s1 and s2 do not have a zero Cj-Zj value.
(iv) The shadow prices of three departments are Rs 33/13 per unit for Department A,
Rs 24/13 per unit for Department B and Rs 0 per unit for Department C.
(v) Dept. A should be given priority to expand the production capacity because it has
the highest shadow price.
11
(vi) If the company produces thirteen units of x3 then 11 units of x1 13 and 8 units
13
8
of x2 13 shall have to be reduced.
13
(vii) If a customer is prepared to pay higher prices for product x3 the price should be
increased by Rs 118/13 so that the company's profit remains unchanged.
(viii) The solution given in the table is not degenerate because the basic variable do not
have zero in the quantity or solution column.
Summary:
For every Linear Programming Problem there exists a Dual that expresses the problem
with a resource orientation. The original problem is called the Primal Problem. The
solution of a Linear Programming Problem can be read from the final simplex tableau of
either Primal or its Dual for all the variables.
At times it is easier to solve the dual problem instead of the Primal Problem, because of
easier computation.
The objective in the Dual is the opposite of the objective in the Primal. If the Primal is
a Maximization Problem then the Dual will be a Minimization Problem and vice versa. The
coefficients of the decision variables in the Objective function of the Primal become
the quantities on the RHS of the resource constraints of the Dual and the vice versa.
The Technological Coefficients in the constraints of the dual problem are the
transpose of the technological coefficients of the constraints in the Primal.
constraints in the Primal correspond to constraints in the Dual and vice versa.
Number of decision variables in the Primal is equal to the number of constraints in
the Dual and vice versa.
In a Standard Linear programming problem, there is non-negativity restriction on
the variables. It implies that the variables can assume values that are greater than or
equal to zero and cannot be negative. When a variable is unrestricted, it can assume
positive, zero or negative value. Therefore, we have to convert them into non-negative
variables by expressing the unrestricted variable as a difference of two non-negative
variables.
The shadow price associated with a particular constraint is the change in the optimal
value of the objective function per unit increase in the right hand-side value for that
constraint, all other problem data remaining unchanged. In the optimal solution, the
objective function coefficients (Zj values) for slack and surplus variables are called the
shadow prices or the dual prices.
A shadow price represents the maximum price at which we should be willing to buy
additional units of a resource. It also represents the minimal price at which we should be
willing to sell units of the resource. A shadow price might therefore be thought of as the
value per unit of a resource.
Existence of a slack/surplus variable in the final simplex tableau indicates that the
corresponding resource has not been fully utilized and its shadow price is zero.
The optimum solution for a dual can be read from the Final Simplex tableau of the Primal
problem and vice versa. The optimal value of objective function is same for both the
Primal and the Dual solutions.
The solution for the decision variables in the Dual can be read from the Final Simplex
Tableau of the Primal Problem. The value of the variable in the Dual corresponding to the
constraint in the Primal is the same as the absolute value of the entry in Cj-Zj row
under the corresponding surplus variable. The solution of the variable in the Dual
corresponding to the = constraint in the Primal is read from the corresponding Zj value
for Artificial variable. The value of the Variable in the Dual corresponding to the <=
constraint in the Primal is read from the absolute value of entry in Cj-Zj row under the
corresponding variable.
The dual variables represent various constraints in the primal problem, therefore their
values in the final simplex solution reflects the shadow prices of the corresponding
resources. The values of dual variables in the optimal solution represent the
corresponding the shadow prices or the worth of the constraints in the Primal. In a
production problem, the shadow prices indicate which of the resource should be given
priority. It is the one with the highest shadow price. The resources with zero shadow
prices are the under-utilized ones and they have spare capacity.
Sensitivity Analysis studies how the optimal solution a linear programming is affected by
the change in its parameters. It helps us to understand the impact of the changes in the
parameters (objective function coefficients, right hand sides and technological
coefficients) of the LPP on the optimal solution.
Glossary:
Artificial variable: Artificial variable is added to the left-hand side of >= and =
constraints.
Basic Solution: It is a solution obtained by setting n-m variables equal to zero and
solving for the remaining m variables (decision, slack, surplus and artificial variables) is
called a basic solution, where n is the number of variables and m is the number of
linear simultaneous equations (constraints) and n>m.
Linear Programming: It is a mathematical technique for optimum allocation of limited
or scarce resources, such as labour, material, machine, money, energy and so on , to
several competing activities such as products, services, jobs and so on, on the basis of a
given criteria of optimality.
Non-basic Variable: The variables in a LPP that are set equal to zero are called non-
basic variables.
Objective function: It is the criterion for selecting the best values of the decision
variables such as profit maximization, sales maximization, cost minimization, etc.
Pivot column: For a non-optimal solution, the column containing the incoming variable
is the Pivot Column. It is the column with the variable having largest value for a
maximization problem and with most negative for a minimization problem as
incoming variable.
Pivot Number: The Pivot number is the number which lies at the intersection of the
Pivot Row and Pivot Column.
Pivot row: The basic variable row with the least non-negative replacement ratio that
is found by dividing each with the corresponding element in the Pivot Column is called
the pivot row.
Slack Variable: It is a non-negative variable added to the left-hand side of type
constraint to make it equal to type constraint. In economic terms the slack variable
represents the unused capacity (requirement minus production).
Surplus variable: It is a non-negative variable subtracted from the left-hand side of
>= type constraint to make it equal to type constraint. In economic terms the
surplus variable represents the excess generated over the requirement (production
minus requirement).
Exercises:
I. Fill in the Blanks:
1. The optimal value of Objective function of a Dual is _________ the optimal
value of Objective function of the associated Primal.
2. The objective in the Dual is the ________of the objective in the Primal.
3. If the Primal is a Minimization Problem then the Dual will be a ___________Problem.
4. The coefficients of the decision variables in the objective function of the Dual are the
same as the _________________ of the resource constraints of the Primal.
5. The Technological Coefficients in the constraints of the dual problem are the
________ of the technological coefficients of the constraints in the Primal.
6. Number of ____________in the Primal is equal to the number of constraints in the
Dual and vice versa.
7. In a Standard Linear programming problem, there is a ______________restriction on
the variables.
8. Dual of a Dual LPP is the __________.
9. The value of the dual variables represent the ____________of the corresponding
resources.
10. ____________ analysis helps us to understand the impact of the changes in the
parameters (objective function coefficients, right hand sides and technological
coefficients) of the LPP on the optimal solution.
Magazine Television
5. Solve the following LPP by simplex method. The company makes three different kinds
of products which are processed in two departments X and Y.
Products
A B C Capacity
Department
X 1 4 5 2000
(hrs.)
Time
Y 2 5 7 2100
Cj 4 8 5 1 0 0 0
(i) Write the optimal product mix and the profit contribution as per the solution given
in the above tableau.
(ii) Is this solution feasible? Why? Give brief reason(s).
(iii) Does the above problem have an alternative optimal solution? If so, show one such
alternate solution.
(iv) Indicate the shadow prices of three machines.
(v) Which of the constraints are binding?
(vi) If the company produces 20 units of x1 how many units of other products shall
have to be reduced, if any?
(vii) If a customer is prepared to pay higher prices for product x4 how much should the
price be increased so that the company's profit remains unchanged?
(viii) Indicate whether the solution given in the table is degenerate. If yes, which
variable is degenerated?
50 x2 1 1 1 0 10
0 s2 1 0 1 1 3
Cj 40 50 0 0
Zj 50 50 50 0 500
(i) Write the optimal product mix and the profit contribution shown by the above
solution.
(ii) Is this solution feasible? Why? Give brief reason(s).
(iii) Does the problem have any alternative optimal solution? If so, show one such
solution.
(iv) If S1 is slack in Machine A and S2 is slack in Machine B, indicate the shadow prices
of the machines.
(v) If the company wishes to expand the production capacity, which of the machines
should be given priority.
(vi) If a customer is prepared to pay higher prices for product x1 how much should the
price be increased so that the company's profit remains unchanged?
(vii) Indicate whether the solution given in the table is degenerate. If yes, which
variable is degenerated?
8. Given below is a simplex tableau for a maximization linear programming problem:
4 5 8 0 0 0
Basic Solution
Variable
5 2 10 0 1 3/5 0 1 0
4 1 5 1 0 20/4 0 -1 0
0 1 20 0 0 -1/5 1 0 1
9. Given below is the final simplex tableau for a maximization linear programming
problem where the variables , , are the units of Product A, B and C
respectively and , , are the slack variables representing Material (in kg.),
Machine (in hrs.) and Labour (in hours).
5 15 8 0 0 0
Basic Solution
Variable
15 2 10 1/3 1 0 1/3 -1/6 0
Solution of Primal:
Minimum costt Z=Rs1000 and 1 = 10 2 = 3.
Thus, 10 advertisements on TV and 3 in magazine will minimize the expenditure.
3. (i) Formulation:
Let x1 and x2 be the number of units of Product A and B respectively. The
required LPP is:
Maximize = 101 + 22 (Total profit)
Subject to the constraints:
101 + 22 35 (Machine Hour constraint)
1 + 0.52 600 (Raw material constraint)
2 800 (Product B constraint)
1 , 2 , 3 0
(ii) The optimum solution is 35 with x1 =3.5 and x2 = 0. Thus 3.5 units of Product A
will be produced by the firm giving a profit of Rs35.
(iii) Dual Problem:
Let 1 , 2 and 3 be dual variables corresponding to the Machine, Raw material
and Product constraints respectively. The Dual can be written as:
Minimize = 351 + 6002 + 8003
Subject to the constraints:
10 1 + 2 10 (Product A constraint)
21 + 0.52 + 3 2 (Product B constraint)
1 , 2 , 3 0
4. Let 1 and 2 represent the number of packets of size I and size II respectively. LPP:
Minimize = 1 + 2
Subject to the constraint
201 + 102 120 (Biscuit A)
501 + 802 740 (Biscuit B)
101 + 602 240 (Biscuit C)
1 , 2 0
Dual Problem:
Let 1 , 2 and 3 be dual variables corresponding to the Machine, Raw material and
Product constraints respectively. The Dual can be written as:
Minimize = 1201 + 7402 + 2403
Subject to the constraints
20 1 + 502 + 103 1 (Size I constraint)
101 + 802 + 603 1 (Size II constraint)
1 , 2 , 3 0
Solution: Minimum number of packets Z=10 is the optimal solution with 2 Size I
packets and 8 Size II packets, given by 1 = 2 2 = 8.
5. (i) Formulation: Let 1 , 2 and 3 be the units of product A, B and C respectively. The
LPP can be expressed as under:
Maximize = 101 + 152 + 123 (Total profit)
Subject to the constraints
1 + 42 + 53 2000 (Dept. X constraint)
21 + 32 + 73 2100 (Dept. Y constraint)
1 , 2 , 3 0
(ii) The optimum solution is 10500 with x1 =380, x2 = 480 and x3 = 0. Product C is
not produced by the firm because it incurs a loss of Rs 23 per unit as shown by the
value in Cj Zj row for 3 to produce.
(iii) There should be Rs 23 per unit increase in profit of Product C for it to be produced.
(iv) Dual Problem:
Let 1 and 2 be dual variables corresponding to constraints of Dept. X and Y
respectively. The Dual can be written as:
Minimize = 20001 + 21002
Subject to the constraints
1 + 22 10 (Product A constraint)
41 + 32 15 (Product B constraint)
51 + 72 12 (Product C constraint)
1 , 2 0
The optimum solution of Dual problem is 10500 with 1 =0 and 2 = 5.
(v) Shadow prices of the Dept. X and Dept. Y are Rs 0 and Rs5 per hour respectively.
6.
4 8 5 1 0 0 0
7 11
825 5.45 8 5 6.5 0
10 5
7 11
1.45 0 0 5.5 0
10 5
(i) The optimal product mix is 1 = 0, 2 = 25 3 = 125 and the profit contribution is Rs
825 by the above solution.
(ii) This solution feasible because there is no artificial variable in the optimal solution.
(iii) No, the problem does not have any alternative optimal solution because the non-
basic variables viz., x1, s1 and s3 do not have a zero Cj-Zj value.
7 11
(iv) The shadow prices of three machines X(s1), Y(s3), Z(s3) is Rs , Rs0 and Rs
10 5
respectively.
(v) The constraints represented by X(s1) and Z(s3) are binding.
1 13
(vi) If the company produces 20 units of x1 then X20=1 unit of product C and X 20
20 20
=13 units of product B shall have to be reduced.
(vii) If a customer is prepared to pay higher prices for product x4 the price be increased
by Rs 5.5 so that the company's profit remains unchanged.
(viii) The solution given in the table is not degenerate because the basic variable do not
have zero in the quantity or solution column.
7.
(i) The optimal product mix is 1 = 0 , 2 = 10 . The profit contribution is Rs
500.
(ii) This solution feasible because there is no artificial variable in the optimal solution.
(iii) No, the problem does not have any alternative optimal solution because the non-
basic variables viz., x1 and s1 do not have a zero Cj-Zj value.
(iv) The shadow prices Machine A and B are Rs 50 per hr. and Rs 0 per hr.
respectively.
(v) Machine A should be given priority to expand the production capacity because it
has the higher shadow price.
(vi) If a customer is prepared to pay higher prices for product x1 the price should be
increased by Rs 10 so that the company's profit remains unchanged.
(vii) The solution given in the table is not degenerate because the basic variable do not
have zero in the quantity or solution column.
8. Simplex tableau-I
4 5 8 0 0 0
Basic Solution
Ratio=
Variable
5 2 10 0 1 3/5 0 1 0 10/0=
4 1 5 1 0 20/4 0 -1 0 5/0=
0 1 20 0 0 -1/5 1 0 1 20/1=20
70 4 5 23 0 1 0
0 0 -15 0 -1 0
4 5 8 0 0 0
Basic Solution
Variable
5 2 10 0 1 3/5 0 1 0
4 1 5 1 0 20/4 0 -1 0
0 3 20 0 0 -1/5 1 0 1
70 4 5 23 0 1 0
0 0 -15 0 -1 0
5 15 8 0 0 0
Basic Solution
Variable
(i) The optimal Production mix is Product A nil, Product B=10 units and Product
C=20 units.
Basic Solution
Variable
2 10 -1/6
3 5 5/12
1
(i) New level of units of Product B to be produced=10+ X 36 =4 units.
6
5
(ii) New level of units of Product C to be produced=5+ X36 =20 units.
12
Basic Solution
Variable
2 10 1/3
3 5 -1/3
1
(i) New level of units of Product B to be produced=10 X 12 =6 units.
3
1
(ii) New level of units of Product C to be produced=5 X36 =17 units.
3
Basic Solution
Variable
2 10 1/3
3 5 2/3
1
(i) New level of units of Product B to be produced=10 X 6 =8 units.
3
2
(ii) New level of units of Product C to be produced=5 X 6 =1 unit.
3
Suggested Readings:
Anthony, M. and N. Biggs. Mathematics for Economics and Finance. Cambridge
University Press.
Ayres, Frank Jr. Theory and Problems of Mathematics of Finance. Schaums Outlines
Series. McGraw Hill Publishing Co.
Budnick, P. Applied Mathematics. McGraw Hill Publishing Co.
Dowling, E.T. Mathematics for Economics, Schaums Outlines Series. McGraw Hill
Publishing Co.
Mizrahi and John Sullivan. Mathematics for Business and Social Sciences. Wiley and
Sons.
Prasad, Bindra and P.K. Mittal. Fundamentals of Business Mathematics. Har-Anand
Publications.
S.K. Sharma and Gurmeet Kaur (2015). Business Mathematics. International Book
House P Ltd. ISBN: 978-93-83283-35-4.
S.K. Sharma and Gurmeet Kaur (2016). Business Mathematics and Statistics.
International Book House P Ltd. ISBN: 978-93-83283-36-8.
Thukral, J.K. Mathematics for Business Studies. Mayur Publications.
Vohra, N.D. Quantitative Techniques in Management. Tata McGraw Hill Publishing Co.
Wikes, F.M. Mathematics for Business, Finance and Economics. Thomson Learning.