Lines, planes, and hyperplanes

The Calculus of Functions Section 1.4
of
Lines, Planes, and Hyperplanes
Several Variables

In this section we will add to our basic geometric understanding of Rn by studying lines
and planes. If we do this carefully, we shall see that working with lines and planes in Rn
is no more diﬃcult than working with them in R2 or R3 .

Lines in Rn
We will start with lines. Recall from Section 1.1 that if v is a nonzero vector in Rn , then,
for any scalar t, tv has the same direction as v when t > 0 and the opposite direction
when t < 0. Hence the set of points

{tv : −∞ < t < ∞}

forms a line through the origin. If we now add a vector p to each of these points, we obtain
the set of points
{tv + p : −∞ < t < ∞},
which is a line through p in the direction of v, as illustrated in Figure 1.4.1 for R2 .

p

v

Figure 1.4.1 A line in R2 through p in the direction of v

Deﬁnition Given a vector p and a nonzero vector v in Rn , the set of all points y in Rn
such that
y = tv + p, (1.4.1)
where −∞ < t < ∞, is called the line through p in the direction of v.

1 Copyright c by Dan Sloughter 2001

2 Lines, Planes, and Hyperplanes Section 1.4

6

4

2 p

-2 -1 1 2 3 4 5
v

-2

-4

Figure 1.4.2 The line through p = (1, 2) in the direction of v = (1, −3)

Equation (1.4.1) is called a vector equation for the line. If we write y = (y1 , y2 , . . . , yn ),
v = (v1 , v2 , . . . , vn ), and p = (p1 , p2 , . . . , pn ), then (1.4.1) may be written as

(y1 , y2 , . . . , yn ) = t(v1 , v2 , . . . , vn ) + (p1 , p2 , . . . , pn ), (1.4.2)

which holds if and only if
y1 = tv1 + p1 ,
y2 = tv2 + p2 ,
. . (1.4.3)
.
. .
.
yn = tvn + pn .
The equations in (1.4.3) are called parametric equations for the line.
Example Suppose L is the line in R2 through p = (1, 2) in the direction of v = (1, −3)
(see Figure 1.4.2). Then

y = t(1, −3) + (1, 2) = (t + 1, −3t + 2)

is a vector equation for L and, if we let y = (x, y),

x = t + 1,
y = −3t + 2

Section 1.4 Lines, Planes, and Hyperplanes 3

5
y

0

10

q
5
p
z
0

-5

-5
0
5
x
Figure 1.4.3 The line through p = (1, 3, 1) and q = (−1, 1, 4)

are parametric equations for L. Note that if we solve for t in both of these equations, we
have
t = x − 1,
2−y
t= .
3
Thus
2−y
x−1= ,
3
and so
y = −3x + 5.
Of course, the latter is just the standard slope-intercept form for the equation of a line in
R2 .
Example Now suppose we wish to ﬁnd an equation for the line L in R3 which passes
through the points p = (1, 3, 1) and q = (−1, 1, 4) (see Figure 1.4.3). We ﬁrst note that
the vector
p − q = (2, 2, −3)
gives the direction of the line, so

y = t(2, 2, −3) + (1, 3, 1)


q-p (q - p) - w

w

p

Figure 1.4.4 Distance from a point q to a line

is a vector equation for L; if we let y = (x, y, z),

x = 2t + 1,
y = 2t + 3,
z = −3t + 1

are parametric equations for L.
As an application of these ideas, consider the problem of finding the shortest distance
from a point q in Rn to a line L with equation y = tv + p. If we let w be the projection
of q − p onto v, then, as we saw in Section 1.2, the vector (q − p) − w is orthogonal to v
and may be pictured with its tail on L and its tip at q. Hence the shortest distance from
q to L is (q − p) − w . See Figure 1.4.4.
Example To find the distance from the point q = (2, 2, 4) to the line L through the
points p = (1, 0, 0) and r = (0, 1, 0), we must first find an equation for L. Since the
direction of L is given by v = r − p = (−1, 1, 0), a vector equation for L is

y = t(−1, 1, 0) + (1, 0, 0).

If we let
v 1
u= = √ (−1, 1, 0),
v 2
then the projection of q − p onto v is

1 1 1
w = ((q − p) · u)u = (1, 2, 4) · √ (−1, 1, 0)) √ (−1, 1, 0) = (−1, 1, 0).
2 2 2

y
0 5 10
-5

4

N M 2
L
0 z

-2

-4
-5
0
5
x 10
Figure 1.4.5 Parallel (L and M ) and perpendicular (L and N ) lines

Thus the distance from q to L is

3 3 82 √
(q − p) − w = , ,4 = = 20.5.
2 2 4

Definition Suppose L and M are lines in Rn with equations y = tv + p and y = tw + q,
respectively. We say L and M are parallel if v and w are parallel. We say L and M are
perpendicular, or orthogonal, if they intersect and v and w are orthogonal.
Note that, by definition, a line is parallel to itself.
Example The lines L and M in R3 with equations
y = t(1, 2, −1) + (4, 1, 2)
and
y = t(−2, −4, 2) + (5, 6, 1),
respectively, are parallel since (−2, −4, 2) = −2(1, 2, −1), that is, the vectors (1, 2, −1) and
(−2, −4, 2) are parallel. See Figure 1.4.5.
Example The lines L and N in R3 with equations
y = t(1, 2, −1) + (4, 1, 2)
and
y = t(3, −1, 1) + (−1, 5, −1),
respectively, are perpendicular since they intersect at (5, 3, 1) (when t = 1 for the first line
and t = 2 for the second line) and (1, 2, −1) and (3, −1, 1) are orthogonal since
(1, 2, −1) · (3, −1, 1) = 3 − 2 − 1 = 0.
See Figure 1.4.5.


Planes in Rn
The following definition is the first step in defining a plane.
Definition Two vectors x and y in Rn are said to be linearly independent if neither one
is a scalar multiple of the other.
Geometrically, x and y are linearly independent if they do not lie on the same line
through the origin. Notice that for any vector x, 0 and x are not linearly independent,
that is, they are linearly dependent, since 0 = 0x.
Definition Given a vector p along with linearly independent vectors v and w, all in Rn ,
the set of all points y such that

y = tv + sw + p, (1.4.4)

where −∞ < t < ∞ and −∞ < s < ∞, is called a plane.
The intuition here is that a plane should be a two dimensional object, which is
guaranteed because of the requirement that v and w are linearly independent. Also
note that if we let y = (y1 , y2 , . . . , yn ), v = (v1 , v2 , . . . , vn ), w = (w1 , w2 , . . . , wn ), and
p = (p1 , p2 , . . . , pn ), then (1.4.4) implies that

y1 = tv1 + sw1 + p1 ,
y2 = tv2 + sw2 + p2 ,
. . (1.4.5)
.
. .
.
yn = tvn + swn + pn .

As with lines, (1.4.4) is a vector equation for the plane and the equations in (1.4.5) are
parametric equations for the plane.
Example Suppose we wish to find an equation for the plane P in R3 which contains the
three points p = (1, 2, 1), q = (−1, 3, 2), and r = (2, 3, −1). The first step is to find two
linearly independent vectors v and w which lie in the plane. Since P must contain the
line segments from p to q and from p to r, we can take

v = q − p = (−2, 1, 1)

and
w = r − p = (1, 1, −2).
Note that v and w are linearly independent, a consequence of p, q, and r not all lying on
the same line. See Figure 1.4.6. We may now write a vector equation for P as

y = t(−2, 1, 1) + s(1, 1, −2) + (1, 2, 1).

Note that y = p when t = 0 and s = 0, y = q when t = 1 and s = 0, and y = r when
t = 0 and s = 1. If we write y = (x, y, z), then, expanding the vector equation,

(x, y, z) = t(−2, 1, 1) + s(1, 1, −2) + (1, 2, 1) = (−2t + s + 1, t + s + 2, t − 2s + 1),

y
1 2 3 4
-1 0

5

v
z

0
w

-5 -2
0
2
x
Figure 1.4.6 The plane y = tv + sw + p, with v = (−2, 1, 1), w = (1, 1, −2), p = (1, 2, 1)

giving us
x = −2t + s + 1,
y = t + s + 2,
z = t − 2s + 1
for parametric equations for P .
To find the shortest distance from a point q to a plane P , we first need to consider the
problem of finding the projection of a vector onto a plane. To begin, consider the plane P
through the origin with equation y = ta + sb where a = 1, b = 1, and a ⊥ b. Given
a vector q not in P , let
r = (q · a)a + (q · b)b,
the sum of the projections of q onto a and onto b. Then

(q − r) · a = q · a − r · a
= q · a − (q · a)(a · a) − (q · b)(b · a)
= q · a − q · a = 0,
2
since a · a = a = 1 and b · a = 0, and, similarly,

(q − r) · b = q · b − r · b
= q · b − (q · a)(a · b) − (q · b)(b · b)
= q · b − q · b = 0.


(q - p) - r

q-p

r

Figure 1.4.7 Distance from a point q to a plane

It follows that for any y = ta + sb in the plane P ,

(q − r) · y = (q − r) · (ta + sb) = t(q − r) · a + s(q − r) · b = 0.

That is, q − r is orthogonal to every vector in the plane P . For this reason, we call r the
projection of q onto the plane P , and we note that the shortest distance from q to P is
q−r .
In the general case, given a point q and a plane P with equation y = tv + sw + p,
we need only ﬁnd vectors a and b such that a ⊥ b, a = 1, b = 1, and the equation
y = ta + sb + p describes the same plane P . You are asked in Problem 29 to verify that
if we let c be the projection of w onto v, then we may take
1
a= v
v

and
1
b= (w − c).
w−c
If r is the sum of the projections of q − p onto a and b, then r is the projection of q − p
onto P and (q − p) − r is the shortest distance from q to P . See Figure 1.4.7.
Example To compute the distance from the point q = (2, 3, 3) to the plane P with
equation
y = t(−2, 1, 0) + s(1, −1, 1) + (−1, 2, 1),
let v = (−2, 1, 0), w = (1, −1, 1), and p = (−1, 2, 1). Then, using the above notation, we
have
1
a = √ (−2, 1, 0),
5
3
c = (w · a)a = − (−2, 1, 0),
5


1
w−c= (−1, −2, 5),
5
and
1
b = √ (−1, −2, 5).
30
Since q − p = (3, 1, 2), the projection of q − p onto P is

1 1
r = ((3, 1, 2) · a)a + ((3, 1, 2) · b)b = −(−2, 1, 0) + (−1, −2, 5) = (11, −8, 5)
6 6

and
1
(q − p) − r = (7, 14, 7).
6
Hence the distance from q to P is
√
294 7
(q − p) − r = =√ .
6 6

More generally, we say vectors v1 , v2 , . . . , vk in Rn are linearly independent if no one
of them can be written as a sum of scalar multiples of the others. Given a vector p and
linearly independent vectors v1 , v2 , . . . , vk , we call the set of all points y such that

y = t1 v1 + t2 v2 + · · · + tk vk + p,

where −∞ < tj < ∞, j = 1, 2, . . . , k, a k-dimensional affine subspace of Rn . In this
terminology, a line is a 1-dimensional affine subspace and a plane is a 2-dimensional affine
subspace. In the following, we will be interested primarily in lines and planes and so will
not develop the details of the more general situation at this time.

Hyperplanes
Consider the set L of all points y = (x, y) in R2 which satisfy the equation

ax + by + d = 0, (1.4.6)

where a, b, and d are scalars with at least one of a and b not being 0. If, for example,
b = 0, then we can solve for y, obtaining

a d
y =− x− . (1.4.7)
b b

If we set x = t, −∞ < t < ∞, then the solutions to (1.4.6) are

a d a d
y = (x, y) = t, − t − = t 1, − + 0, − . (1.4.8)
b b b b


L

n
y y-p

p

Figure 1.4.8 L is the set of points y for which y − p is orthogonal to n

Thus L is a line through 0, − d in the direction of 1, − a . A similar calculation shows
b b
d
that if a = 0, then we can describe L as the line through − a , 0 in the direction of
− a , 1 . Hence in either case L is a line in R2 .
b

Now let n = (a, b) and note that (1.4.6) is equivalent to

n · y + d = 0. (1.4.9)

Moreover, if p = (p1 , p2 ) is a point on L, then

n · p + d = 0, (1.4.10)

which implies that d = −n · p. Thus we may write (1.4.9) as

n · y − n · p = 0,

and so we see that (1.4.6) is equivalent to the equation

n · (y − p) = 0. (1.4.11)

Equation (1.4.11) is a normal equation for the line L and n is a normal vector for L. In
words, (1.4.11) says that the line L consists of all points in R2 whose diﬀerence with p is
orthogonal to n. See Figure 1.4.8.
Example Suppose L is a line in R2 with equation

2x + 3y = 1.


Then a normal vector for L is n = (2, 3); to ﬁnd a point on L, we note that when x = 2,
y = −1, so p = (2, −1) is a point on L. Thus

(2, 3) · ((x, y) − (2, −1)) = 0,

or, equivalently,
(2, 3) · (x − 2, y + 1) = 0,
is a normal equation for L. Since q = (−1, 1) is also a point on L, L has direction
q − p = (−3, 2). Thus
y = t(−3, 2) + (2, −1)
is a vector equation for L. Note that

n · (q − p) = (2, 3) · (−3, 2) = 0,

so n is orthogonal to q − p.
Example If L is a line in R2 through p = (2, 3) in the direction of v = (−1, 2), then
n = (2, 1) is a normal vector for L since v · n = 0. Thus

(2, 1) · (x − 2, y − 3) = 0

is a normal equation for L. Multiplying this out, we have

2(x − 2) + (y − 3) = 0;

that is, L consists of all points (x, y) in R2 which satisfy

2x + y = 7.

Now consider the case where P is the set of all points y = (x, y, z) in R3 that satisfy
the equation
ax + by + cz + d = 0, (1.4.12)
where a, b, c, and d are scalars with at least one of a, b, and c not being 0. If for example,
a = 0, then we may solve for x to obtain

b c d
x=− y− z− . (1.4.13)
a a a
If we set y = t, −∞ < t < ∞, and z = s, −∞ < s < ∞, the solutions to (1.4.12) are

y = (x, y, z)
b c d
= − t − s − , t, s
a a a (1.4.14)
b c d
= t − , 1, 0 + s − , 0, 1 + − , 0, 0 .
a a a


n

y-p

P

Figure 1.4.9 P is the set of points y for which y − p is orthogonal to n

Thus we see that P is a plane in R3 . In analogy with the case of lines in R2 , if we let
n = (a, b, c) and let p = (p1 , p2 , p3 ) be a point on P , then we have

n · p + d = ax + by + cz + d = 0,

from which we see that n · p = −d, and so we may write (1.4.12) as

n · (y − p) = 0. (1.4.15)

We call (1.4.15) a normal equation for P and we call n a normal vector for P . In words,
(1.4.15) says that the plane P consists of all points in R3 whose diﬀerence with p is
orthogonal to n. See Figure 1.4.9.
Example Let P be the plane in R3 with vector equation

y = t(2, 2, −1) + s(−1, 2, 1) + (1, 1, 2).

If we let v = (2, 2, −1) and w = (−1, 2, 1), then

n = v × w = (4, −1, 6)

is orthogonal to both v and w. Now if y is on P , then

y = tv + sw + p

for some scalars t and s, from which we see that

n · (y − p) = n · (tv + sw) = t(n · v) + s(n · w) = 0 + 0 = 0.

That is, n is a normal vector for P . So, letting y = (x, y, z),

(4, −1, 6) · (x − 1, y − 1, z − 2) = 0 (1.4.16)


is a normal equation for P . Multiplying (1.4.16) out, we see that P consists of all points
(x, y, z) in R3 which satisfy
4x − y + 6z = 15.

Example Suppose p = (1, 2, 1), q = (−2, −1, 3), and r = (2, −3, −1) are three points
on a plane P in R3 . Then
v = q − p = (−3, −3, 2)
and
w = r − p = (1, −5, −2)
are vectors lying on P . Thus

n = v × w = (16, −4, 18)

is a normal vector for P . Hence

(16, −4, 18) · (x − 1, y − 2, z − 1) = 0

is a normal equation for P . Thus P is the set of all points (x, y, z) in R3 satisfying

16x − 4y + 18y = 26.

The following definition generalizes the ideas in the previous examples.
Definition Suppose n and p are vectors in Rn with n = 0. The set of all vectors y in
Rn which satisfy the equation
n · (y − p) = 0 (1.4.17)
is called a hyperplane through the point p. We call n a normal vector for the hyperplane
and we call (1.4.17) a normal equation for the hyperplane.
In this terminology, a line in R2 is a hyperplane and a plane in R3 is a hyperplane. In
general, a hyperplane in Rn is an (n − 1)-dimensional affine subspace of Rn . Also, note
that if we let n = (a1 , a2 , . . . , an ), p = (p1 , p2 , . . . , pn ), and y = (y1 , y2 , . . . , yn ), then we
may write (1.4.17) as

a1 (y1 − p1 ) + a2 (y2 − p2 ) + · · · + an (yn − pn ) = 0, (1.4.18)

or
a1 y1 + a2 y2 + · · · + an yn + d = 0 (1.4.19)
where d = −n · p.
Example The set of all points (w, x, y, z) in R4 which satisfy

3w − x + 4y + 2z = 5

is a 3-dimensional hyperplane with normal vector n = (3, −1, 4, 2).


n
q
H
q-p

| (q - p) . u|

p

Figure 1.4.10 Distance from a point q to a hyperplane H

The normal equation description of a hyperplane simplifies a number of geometric
calculations. For example, given a hyperplane H through p with normal vector n and a
point q in Rn , the distance from q to H is simply the length of the projection of q − p
onto n. Thus if u is the direction of n, then the distance from q to H is |(q − p) · u|. See
Figure 1.4.10. Moreover, if we let d = −p · n as in (1.4.19), then we have
q·n−p·n |q · n + d|
|(q − p) · u| = |q · u − p · u| = = . (1.4.20)
n n
Note that, in particular, (1.4.20) may be used to find the distance from a point to a line
in R2 and from a point to a plane in R3 .
Example To find the distance from the point q = (2, 3, 3) to the plane P in R3 with
equation
x + 2y + z = 4,
we first note that n = (1, 2, 1) is a normal vector for P . Using (1.4.20) with d = −4, we
see that the distance from q to P is
|q · n + d| |(2, 3, 3) · (1, 2, 1) − 4| 7
= √ =√ .
n 6 6
Note that this agrees with an earlier example.
We will close this section with a few words about angles between hyperplanes. Note
that a hyperplane does not have a unique normal vector. In particular, if n is a normal
vector for a hyperplane H, then −n is also a normal vector for H. Hence it is always
possible to choose the normal vectors required in the following definition.
Definition Let G and H be hyperplanes in Rn with normal equations

m · (y − p) = 0


and
n · (y − q) = 0,
respectively, chosen so that m · n ≥ 0. Then the angle between G and H is the angle
between m and n. Moreover, we will say that G and H are orthogonal if m and n are
orthogonal and we will say G and H are parallel if m and n are parallel.
The effect of the choice of normal vectors in the definition is to make the angle between
the two hyperplanes be between 0 and π .2

Example To find the angle θ between the two planes in R3 with equations

x + 2y − z = 3

and
x − 3y − z = 5,
we first note that the corresponding normal vectors are m = (1, 2, −1) and n = (1, −3, −1).
Since m · n = −4, we will compute the angle between m and −n. Hence

m · (−n) 4 4
cos(θ) = =√ √ =√ .
m n 6 11 66

Thus, rounding to four decimal places,

4
θ = cos−1 √ = 1.0560.
66

See Figure 1.4.11.
Example The planes in R3 with equations

3x + y − 2z = 3

and
6x + 2y − 4z = 13
are parallel since their normal vectors are m = (3, 1, −2) and n = (6, 2, −4) and n = 2m.

Problems

1. Find vector and parametric equations for the line in R2 through p = (2, 3) in the
direction of v = (1, −2).
2. Find vector and parametric equations for the line in R4 through p = (1, −1, 2, 3) in
the direction of v = (−2, 3, −4, 1).
3. Find vector and parametric equations for the lines passing through the following pairs
of points.


y

1 0 -1 -2

0

z

-5

-10
2
1
0
-1
-2
x

Figure 1.4.11 The planes x + 2y − z = 3 and x − 3y − z = 5

(a) p = (−1, −3), q = (4, 2) (b) p = (2, 1, 3), q = (−1, 2, 1)
(c) p = (3, 2, 1, 4), q = (2, 0, 4, 1) (d) p = (4, −3, 2), q = (1, −2, 4)

4. Find the distance from the point q = (1, 3) to the line with vector equation y =
t(2, 1) + (3, 1).


5. Find the distance from the point q = (1, 3, −2) to the line with vector equation y =
t(2, −1, 4) + (1, −2, −1).
6. Find the distance from the point r = (−1, 2, −3) to the line through the points p =
(1, 0, 1) and q = (0, 2, −1).
7. Find the distance from the point r = (−1, −2, 2, 4) to the line through the points
p = (2, 1, 1, 2) and q = (1, 2, −4, 3).
8. Find vector and parametric equations for the plane in R3 which contains the points
p = (1, 3, −1), q = (−2, 1, 1), and r = (2, −3, 2).
9. Find vector and parametric equations for the plane in R4 which contains the points
p = (2, −3, 4, −1), q = (−1, 3, 2, −4), and r = (2, −1, 2, 1).
10. Let P be the plane in R3 with vector equation y = t(1, 2, 1) + s(−2, 1, 3) + (1, 0, 1).
Find the distance from the point q = (1, 3, 1) to P .
11. Let P be the plane in R4 with vector equation y = t(1, −2, 1, 4)+s(2, 1, 2, 3)+(1, 0, 1, 0).
Find the distance from the point q = (1, 3, 1, 3) to P .
12. Find a normal vector and a normal equation for the line in R2 with vector equation
y = t(1, 2) + (1, −1).
13. Find a normal vector and a normal equation for the line in R2 with vector equation
y = t(0, 1) + (2, 0).
14. Find a normal vector and a normal equation for the plane in R3 with vector equation
y = t(1, 2, 1) + s(3, 1, −1) + (1, −1, 1).
15. Find a normal vector and a normal equation for the line in R2 which passes through
the points p = (3, 2) and q = (−1, 3).
16. Find a normal vector and a normal equation for the plane in R3 which passes through
the points p = (1, 2, −1), q = (−1, 3, 1), and r = (2, −2, 2).
17. Find the distance from the point q = (3, 2) in R2 to the line with equation x+2y−3 = 0.
18. Find the distance from the point q = (1, 2, −1) in R3 to the plane with equation
x + 2y − 3x = 4.
19. Find the distance from the point q = (3, 2, 1, 1) in R4 to the hyperplane with equation
3x + y − 2z + 3w = 15.
20. Find the angle between the lines in R2 with equations 3x + y = 4 and x − y = 5.
21. Find the angle between the planes in R3 with equations 3x−y +2z = 5 and x−2y +z =
4.
22. Find the angle between the hyperplanes in R4 with equations w + x + y − z = 3 and
2w − x + 2y + z = 6.
23. Find an equation for a plane in R3 orthogonal to the plane with equation x+2y−3z = 4
and passing through the point p = (1, −1, 2).


24. Find an equation for the plane in R3 which is parallel to the plane x − y + 2z = 6 and
passes through the point p = (2, 1, 2).
25. Show that if x, y, and z are vectors in Rn with x ⊥ y and x ⊥ z, then x ⊥ (ay + bz)
for any scalars a and b.
26. Find parametric equations for the line of intersection of the planes in R3 with equations
x + 2y − 6z = 4 and 2x − y + z = 2.
27. Find parametric equations for the plane of intersection of the hyperplanes in R4 with
equations w − x + y + z = 3 and 2w + 4x − y + 2z = 8.
28. Let L be the line in R3 with vector equation y = t(1, 2, −1) + (3, 2, 1) and let P be the
plane in R3 with equation x + 2y − 3z = 8. Find the point where L intersects P .
29. Let P be the plane in Rn with vector equation y = tv+sw +p. Let c be the projection
of w onto v,
1
a= v,
v
and
1
b= (w − c).
w−c
Show that y = ta + sb + p is also a vector equation for P .

Lines, planes, and hyperplanes

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Lines, planes, and hyperplanes (20)

More from Tarun Gehlot (20)

Lines, planes, and hyperplanes