Professional Documents
Culture Documents
Contents
0.1
~v
0.1.1
0.1.2
P1
0.1.3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0.1.4
0.1.5
0.1.6
0.1.7
0.1.8
0.1.9
and
P2
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pl
and Plane
~n
Through Point
~v1
P1 , P2 ,
P l1
and
P l1
Pl
P3
and
. .
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
P l2
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
P l2
and
to Plane
~v2
and
. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .
{~v , w}
~ ,
normal vectors
{x, y, z}, equation coecients {a, b, c, d}, lines {l}, points {P, Q},
{~n}, line-parameters {s, t}, planes {P l}, angles {, , }.
~v1
and
~v2
forming an angle
~v1 ~v2
w
~
onto
~v
we have
arbitrary vectors
is orthogonal to both
is given by
w
~
in the direction
~v .
P~1
and
P~2
is given by
~v = P~2 P~1 .
a nonzero scalar.
Two lines are parallel if their direction vectors are (nonzero) scalar multiples of one another.
ax + by + cz = d
~n = ha, b, ci
The normal vector is orthogonal to every line that lies in the plane. It is unique up to multiplication by
a nonzero scalar.
0.1.1
~v
~v = h1, 2, 1i
l1
has equation
l : hx, y, zi = P~ + t ~v
P = (1, 2, 2).
~v = h6, 6, 6i
P = (4, 1, 100).
l2 : hx, y, zi =
h1, t 1, 5t + 2i.
0.1.2
h0, 1, 5i.
P1
and
P2 ,
(0, 0, 0)
and
We compute ~
v
(1, 1, 3)
~v = P~2 P~1 ,
(1, 2, 3).
~v = h1 0, 2 0, 3 0i = h1, 2, 3i.
0.1.3
l2 : hx, y, zi = h2, t + 2, 5t + 3i
Since l1 and l2 are parallel we see that the direction vector for l1
l : hx, y, zi = P~1 + t ~v .
h0, 1, 5i.
and
(2, 2, 2).
l : hx, y, zi = ht + 1, t 1, t + 3i
Note 1: If the resulting equation has no solution, the line and plane do not intersect. Equivalently, the
line is parallel to the plane.
Note 2: If the resulting equation reduces to an identity (i.e., is always true) then every point on the line
lies inside the plane. Equivalently, the line itself lies inside the plane.
t = 2,
which is just
x = 2t + 1, y = t 1, z = t.
5 = (2t + 1) + 2(t 1) + 3(t) = 3t 1
l : hx, y, zi = h2t + 1, t 1, ti
from which
(x, y, z) = (5, 3, 2)
x + 2y + 3z = 5.
t = 2.
5 + 2(3) + 3(2) = 5.
x = 3t 1, y = 2t + 2, z = 4t.
the plane gives 1 = 2(2t + 2) + (4t) = 4
2y + z = 1.
1 = 4,
any solution.
Hence the line and plane do not intersect, which is to say, the line is parallel to the plane.
x = 3t 1, y = 2t + 2, z = 4t.
4 = 2(2t + 2) + (4t) = 4
4 = 4,
2y + z = 4.
which is an identity.
Hence every point on the line lies in the plane, which is to say, the line lies inside the plane.
0.1.4
solution, it will give the intersection point. If there is no solution, the lines do not intersect.
3t = 2s + 1.
2s + 4 = s + 1
t = 1, and
h1, 2, 3i on l1 .
Back-subbing gives
gives the point
ht, 2t, 3ti = hs + 2, s + 1, 2s + 1i, which is just the 3 equations t = s+2, 2t = s+1,
we check that
intersect at
(1, 2, 3)
{s = 1, t = 1}
so that
s = 1.
t=1
l1 : hx, y, zi = h2t + 1, t 1, ti
and
l2 : hx, y, zi =
hs + 2, s + 1, 2s + 1i.
0.1.5
Back-subbing gives
Idea: Given
However,
{s = 7, t = 3}
t 1 = 2t + 2
t = 3.
hence
do not intersect .
~n = ha, b, ci
(1, 1, 1)
We plug in
(1, 1, 1)
to see
1+21=d
so
x + 2y z = d
(2, 3, 2)
We plug in
to see
d = 3,
(2, 3, 2)
ax + by + cz =
.
0.1.6
s = 7.
To nd
s = 2t + 1,
12 18 + 12 = d
~n = h1, 2, 1i.
6x + 6y + 6z = d
so
d = 6,
d.
x + 2y z = 3
~n = h6, 6, 6i.
d.
6x + 6y + 6z = 6
x + y + z = 1.
~v1 ~v2
~v1
and
~v2 ,
~v1
and
~v2
~n = ~v1 ~v2 .
Variation 1: If instead the plane is asked to contain be parallel to two arbitrary lines, we can reduce to
the above. We need the plane's normal vector to be orthogonal to both lines, and so we take
where
~v1
and
~v2
~n = ~v1 ~v2
Note: This setup includes the special case where the lines l1 and l2 intersect. In this case we will be
nding a plane parallel to the plane containing l1 and l2 .
Variation 2: If instead we ask for a plane containing one line l1 and parallel to another line l2 , we again
compute
~n = ~v1 ~v2 ,
We compute
~v1 = h1, 1, 1i
and
~v2 = h1, 2, 3i
and passing
(0, 0, 0).
~n = ~v1 ~v2 =
~i ~j ~k
1
1 1 1 =
2
1 2 3
1 ~ 1
i
1
3
~j + 1
1
1
3
1 ~
k = h1, 2, 1i.
2
and
(1)(3) = 0.
This assures us that the normal vector we found really is orthogonal to both vectors, which we know
it's supposed to be.
We compute
x 2y + z = 0
for some
d.
d = 0.
h2t + 1, t 1, ti
We have
and l2
~v1 = h2, 1, 1i
We compute
x 2y + z = d
and
~n = ~v1 ~v2 =
Sanity check:
We compute
1 ~ 2
i
2
1
3x + 5y z = d
3x + 5y z = 4
(1, 1, 2),
and
for some
so that
1 ~ 2
j+
2
1
1 ~
k = h3, 5, 1i.
1
d.
We computed the cross product of the direction vectors of l1 and l2 in the previous example; we obtained
h3, 5, 1i.
0.1.7
Plugging in gives
We see in particular that the lines l1 and l2 cannot intersect, because they lie in dierent parallel planes.
for some
t = 0.
d.
(1, 1, 0).
d = 3(1) + 5(1) 0 = 2.
3x + 5y z = 2
s=0
to get
(2, 1, 1).
~n.
By the previous example we just need to nd two vectors lying in the
3x + 5y z = d,
~n.
We have
Then
and
and
Sanity check:
P~1
to see
7x 5y + 8z = 13
and
Then
Sanity check:
and
P~1
to see
and
7(3) 5(0) +
P~3 = (0, 1, 0).
7x + 2y + 3z = d,
for some
d.
For an extra error check, we verify that all three points do lie in this plane: we have
and
7(2) + 2(2) +
P l1
and
P l2
(x, y, z)
We have
~n1 = h1, 1, 2i
Sanity check:
and
Solving yields
~n1 = h2, 1, 4i
Sanity check:
and
2x + y z = 0.
and
and
so
(0, 3, 3)
x = 0:
this requires
y + 2z = 3
and
l : hx, y, zi = ht, 5t + 3, 3t + 3i
2x + y 4z = 7
and
3x 4y + 5z = 2.
4y + z = 2.
x y + 2z = 3
y = z = 3,
and
~n1
y z = 0.
d.
~v1 = P~2 P~1 = h3, 3, 5i and ~v2 = P~3 P~1 = h2, 1, 4i.
3 5
~i 3 5 ~j + 3 3 ~k = h7, 2, 3i.
~n = ~v1 ~v2 =
2 1
2 4
1 4
We have
for some
For an extra error check, we verify that all three points do lie in this plane: we have
0.1.8
7x 5y + 8z = d,
and
Solving yields
y = 1
and
and
z = 2,
so
(0, 1, 2)
x = 0:
this requires
y 4z = 7
and
l : hx, y, zi = h11t, 2t 1, 5t 2i
0.1.9
= ,
N1
and
N2
where
0 .
2
We have
Hence
~n1 = h1, 1, 2i
and
We have
x y + 2z = 3
and
with
and
~n1 ~n2 = 30
along with
|~n1 ~n2 |
30
Thus the Dot Product Theorem gives cos() =
= =
||~n1 || ||~n2 ||
5 42
!
42
Hence the angle between the planes is cos1
radians .
7
6.
3x 4y + 5z = 2.
5 2.
0.1.10
|~n1 ~n2 |
||~n1 || ||~n2 ||
radians .
2x + y 4z = 7
cos() =
2x + y + z = 0.
with
~n1 = h2, 1, 4i
and
||~n1 || =
21
and
||~n2 || =
42
.
7
the plane. In theory we could use this information to parametrize the line joining
intersection of the line and plane, and nally compute the distance we seek. However, there is a less messy
way using vector projections:
Clever Idea:
Let
~n
P l.
on the plane
~v
connecting
normal
~n.
Explicitly,
~v = Proj~n (w)
~ =
~v
is
||~v || =
|~n w|
~
||~n||
~ =d
~n Q
and
(x0 , y0 , z0 )
x+y+z =1
a2 + b2 + c2
to the origin
we compute
ax + by + cz = d.
onto the
~n.
to the plane
~ P~
w
~ =Q
~n w
~
~n ~n
P l.
(0, 0, 0).
~n w
~ =
P l.
~n = h1, 1, 1i, along with P = h0, 0, 0i. We also grab any old
compute w
~ = h1, 0, 0i, and then get ~n w
~ = 1 and ||~n|| = 3.
We
We have
1
||~v || =
3
We have
~n = h2, 1, 2i,
along with
point
Q = h1, 0, 0i
on the plane.
=
2
2
2
3
1 +1 +1
2x y + 2z = 12
P = h2, 1, 3i.
to the point
(2, 1, 3).
Q = h6, 0, 0i
on the
plane.
0.1.11
We compute
w
~ = h4, 1, 3i,
~n w
~ =3
||~v || =
3
=1
3
and
||~n|| = 3.
Idea: Imagine the two lines l1 and l2 as sitting inside two parallel planes
P l1
and
P l2 .
The distance between the lines will then be the same as the distance between the planes, which we can
compute just by nding the distance of any arbitrary point
of
P l1 ,
to the plane
P l2 .
[This we know
In order to nd these two parallel planes, we just need to nd their (common) normal vector
~n,
which
will be orthogonal to the direction vectors of both lines. [This we also know how to do using the cross
product.]
To save us a little work, if we nd the equations for the two planes as
P l2 : ax + by + cz = d2 ,
|d d1 |
2
.
a2 + b2 + c2
We have
~v1 = h2, 1, 1i
and
l1 : hx, y, zi = h2t + 1, t 1, ti
Sanity check:
and a
P l1
and
P l2
and
~n = h3, 5, 1i.
~n ~v2 = (3)(1) + (5)(1) + (1)(2) = 0.
3x + 5y z = d for some d.
t = 0. This gives
P l1 : 3x + 5y z = 2 .
are
d = 3(1) + 5(1) 0 = 2,
and
: hx, y, zi = hs + 2, s + 1, 2s + 1i.
We plug to get
P l1 : ax + by + cz = d1
we can compute using the point-to-plane formula that the distance will be
l2 ,
so we get
l2 : take s = 0
P l2 : 3x + 5y z = 10 .
we need a point on
so we get
to get
(2, 1, 1).
the point
(1, 1, 0).
Plugging in gives
Applying the formula above shows that the distance between these planes, and hence the lines, is
|10 (2)|
p
32
52
(1)2
12
=
35
Remark: If we tried to nd the minimum distance as a regular optimization problem, we would have
it's not easy either way! But our not-so-dicult vector computation shows that the minimum of this
expression is
144
.
35