Vous êtes sur la page 1sur 12

Chapter 1

Introduction to multivariable
functions and vectors

1.3 General vectors, dot products and planes


R, R2 and R3 are nice because they can be visualized, but it is still natural
to consider vectors in higher dimensions. Indeed, some mathematical models
involve more than 3 variables. For completeness we provide the following:
Definition 1.3.1 (General vectors, addition, scalar multiplication, magni-
tude). A vector in Rn is an array of n number ~v = hv1 , . . . , vn i in angled
brackets. Given another vector w ~ = hw1 , . . . , wn i and a scalar α ∈ R we have
the following addition and scalar multiplication formulas:

~v + w~ = hv1 + w1 , . . . , vn + wn i (1.3.1)
α · ~v = hαv1 , . . . , αvn i . (1.3.2)

Although higher dimension geometry seems far-fetched we also provide


the following:
Definition 1.3.2. Given points P = (p1 , . . . , pn ), Q = (q1 , . . . , qn ) ∈ Rn the
displacement vector from P to Q is

P~Q = hq1 − p1 , . . . , qn − pn i = hQ − P i . (1.3.3)

The magnitude of a vector ~v is given by the Pythagorean formula:


q
|~v | = v12 + . . . + vn2 . (1.3.4)

1
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 2

And the distance between two points is by the formula:



~
d(P, Q) = P Q . (1.3.5)

The remarkable dot product


Geometry is all about angles and distances. The dot product completely
defines what we know as the Euclidean geometry on Rn . Because we consider
vectors as arrows that are free to float around Rn it makes sense to try to
give a precise definition of the angle between two vectors.
Definition 1.3.3. Let ~v and w ~ be vectors in Rn then the angle between ~v
and w~ is the angle θ shown below with the prescribed side lengths

| |~v −
|w~ w
~|

θ
|~v |

Which can be calculated using the Law of cosines to be θ such that

~ 2 = |w|
|~v − w| ~ 2 + |~v |2 − 2|~v ||w|
~ cos(θ) (1.3.6)

In particular if you can calculate all those magnitudes you can solve for
cos(θ) and then arccos it. Informally it’s the angle between the vectors when
you put their backsides together. Doing this in 2D is straightforward.
Example 1.3.1. What is the angle between the vectors w ~ = h−5, 4i and
~v = h2, 2i?

h−5, 4i y
w~
h2, 2i
θ ~u
x
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 3

Previously, by normalizing and using arccos we found the angle θ to be about


141◦ . The angle between h2, 2i originating at the origin (0, 0) and the x-axis
is 45◦ ; so the angle between the two vectors is about 141◦ − 45◦ = 96◦ .
In R3 the angle between two vectors can be found similarly by putting
them end-to-end and measuring the angle in the plane that contains them
both. This is really hard to do, and it’s why the dot product is great.

Definition 1.3.4 (Dot product (algebraic)). Let ~v = hv1 , . . . , vn i and w ~ =


hw1 , . . . , wn i be two vectors, then their dot product is the following scalar.

~v · w
~ = v1 w1 + . . . + v1 wn . (1.3.7)

Although this operation looks silly, its importance cannot be overstated.


At the very least it gives a unified formula for magnitude:

|~v |2 = ~v · ~v (1.3.8)

which in turn defines distance.


Example 1.3.2. Let ~v = h1, 2i and let w
~ = h3, 4i find ~v · w
~ and |w|.
~
Solution.

~v · w
~ = 1 · 3 + 2 · 4 = 3 + 8 = 11
|~v |2 = ~v · ~v = 1 · 1 + 2 · 2 = 5

⇒ |~v | = 5
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 4

Dot products also have the following algebraic properties.


~ be vectors in Rn and let α, β, γ be scalars.
Proposition 1.3.5. Let ~u, ~v , w
Then the following properties hold.

~u · ~v = ~v · ~u (1.3.9)
(~u + ~v ) · w~ = ~u · w ~ + ~v · w
~ (1.3.10)
(α~v ) · w~ = α(~v · w)~ (1.3.11)

The identity (1.3.9) is known as commutativity and all three identities


combine to give a property known as bilinearity.
Using these algebraic properties we can compute the dot product by
“breaking down” into standard basis vectors.
Example 1.3.3. Compute ~u · ~v if ~u = ~i − 2~j + 3~k and ~v = −~j + ~k.
Solution. Using the properties of the dot product we get

~u · ~v = (~i − 2~j + 3~k) · (−~j + ~k)

= ~i · (−~j) + ~i · ~k + (−2~j) · (−~j) + (−2~j) · ~k + (3~k) · (−~j) + (3~k) · ~k


= −(~i · ~j) + ~i · ~k + 2(~j · ~j) − 2(~j · ~k) − 3(~k · ~j) + 3(~k · ~k)
= −1 · 0 + 0 + 2 · 1 − 2 · 0 − 3 · 0 + 3 · 1 = 5
Theorem 1.3.6 (The geometric definition of the dot product). For any two
vectors ~v , w
~ we have
~v · w
~ = |~v ||w|
~ cos(θ) (1.3.12)
where θ is the angle between ~v and w.
~ In particular,

• if ~v and w
~ are orthogonal (or perpendicular) then ~v · w
~ = 0, and

• if ~v and w
~ are parallel then ~v · w
~ = ±|~v ||w|
~

Example 1.3.4. Find the angle θ between ~u and ~v if

(a) ~u = h−1, 2, −3i and ~v = h1, −1, −1i,

(b) ~u = h0, 2, −1i and ~v = h3, 2, 1i.

Solution.
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 5

(a) We have
~u · ~v = (−1) · 1 + 2 · (−1) + (−3) · (−1) = 0
π
which implies that ~u ⊥ ~v , that is, θ = .
2
(b) We have ~u · ~v = 0 · 3 + 2 · 2 + (−1) · 1 = 3. Next,
p √ √ √
|~u| = 02 + 22 + (−1)2 = 5, |~v | = 32 + 22 + 12 = 14

hence, √ √
3= 5 · 14 · cos θ
from which we obtain
 
3 3
cos θ = √ , and θ = arccos √ ≈ 69◦
70 70

Example 1.3.5. Use the properties of the dot product to expand

~ 2 = (~v − w)
|~v − w| ~ · (~v − w)
~

to prove the equality (1.3.12), i.e. the geometric definition of the dot product.

Solution.

~2=
|~v − w|
(~v − w)
~ · (~v − w)
~ = ~v · (~v − w)
~ −w
~ · (~v − w)
~
= ~v · ~v − ~v · w
~ −w
~ · ~v + w ~ = |~v |2 − 2(~v · w)
~ ·w ~2
~ + |w|
~ 2 = |~v |2 + |w|
⇒ |~v − w| ~ 2 − 2(~v · w)
~

So by Definition 1.3.3 of the angle between two vectors we have ~v · w


~ =
|~v ||w|
~ cos(θ).

1.3.1 Normal vectors and planes


Because the dot product detects perpendicularity, it can be used to represent
planes in R3 . Let P = (x0 , y0 , z0 ) be a point in R3 and let ~n = ha, b, ci be a
vector. Then the plane Π passing through P with normal vector ~n consists
of all those points X = (x, y, z) such that ~n · P~X = 0.
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 6

~n

Y
P~Y
P~X X
P

Definition 1.3.7 (Point-normal equation for a plane). If ~n = ha, b, ci is a


vector and P = (x0 , y0 , z0 ) is point then the plane Π with normal vector ~n
and containing P is the set of points X = (x, y, z) that satisfy the following
equivalent equations:

~n · P~X = 0 (1.3.13)
⇔ a · (x − x0 ) + b · (y − y0 ) + c · (z − z0 ) = 0 (1.3.14)
⇔ ax + by + cz = ax0 + by0 + cz0 (1.3.15)
⇔ ~n · X ~ = ~n · P~ (1.3.16)

where P~ = hx0 , y0 , z0 i and X


~ = hx, y, zi are vectorfications of the points P
and X.
You may ask yourself, which of these equations is the best. The answer
is that they all come in handy at different times and if you’re comfortable
with the algebra, then switching between them is natural.
Example 1.3.6. Is the point Q = (1, 2, 3) in the plane Π containing the
point P = (4, 5, 6) with normal vector ~n = h7, 8, 9i?
Solution. We get the plane equation:

~n · hx, y, zi = ~n · P~
⇔ h7, 8, 9i · hx, y, zi = h7, 8, 9i · h4, 5, 6i
⇔ 7x + 8y + 9z = 7 · 4 + 8 · 5 + 9 · 6 = 122

now if Q is a point on that plane then its entries, substituted into x, y, z must
satisfy this equation. We verify
~ = 7 · 1 + 8 · 2 + 9 · 3 = 50 6= 122.
~n · Q

So Q is not on Π.
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 7

Example 1.3.7. You have a plane given by Π : x + 2y + 3z = 4. Find a


normal vector and a point on Π.
Solution. From such an equation we can “read off” the normal vector ~n =
h1, 2, 3i. To find a point we can set y, z = 0 which leaves x = 4 so we take
P = (4, 0, 0). It’s easy to check that this ~n and P give us the equation for Π.
Sometimes there are multiple ways to express a plane.
Example 1.3.8. Consider the point P = (1, −1, 1) and the normal vectors
~n = h−2, 3, −1i and m ~ = h4, −6, 2i these vectors are parallel so they must
define the same plane, however the plane equations we get are:
~ = ~n · P~
~n · X
−2x + 3y − z = −6

and

m~ ·X~ =m~ · P~
4x − 6y + 2z = 12

But note that

4x − 6y + 2z = 12
−1 −1
⇒ (4x − 6y + 2z) = · 12
2 2
−2x + 3y − z = −6

So that the two equations, although they look different, are in fact equivalent,
i.e. they have the same solutions.
It turns out that the point normal equations work in every Rn , and they
give rise to a codimension 1 hyperplane (i.e. an n − 1 dimensional linear
affine subspace). Let’s do the easy to draw case:
Example 1.3.9. Let P = (2, 1) and let ~n h−3, 1i be in R2 . What is the line
line ` passing through P with normal vector ~n?
Solution. Let’s draw a picture
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 8

`
~n
P
x

The point normal formula gives us:

~n · hx, yi = ~n · P~
−3x + y = −3 · 2 + 1 · 1 = −5.

If we isolate y we get the familiar slope-intercept form:y = 3x − 5.


So in a certain sense the normal vector defines a sort of slope.
Definition 1.3.8 (Parallel). Two planes Π1 , Π2 are parallel if they have
parallel normal vectors.
Example 1.3.10. Are the plane Π1 : x+2y+3z = 4 and Π2 : 5x+6y+7z = 8
parallel? If not find a point P that is contained in both of them.
Solution. From the formulas we “read off” normal vectors ~n1 = h1, 2, 3i and
~n2 = h5, 6, 7i. Let’s verify if they are parallel set

α · h1, 2, 3i = h5, 6, 7i

equating first entries forces α = 5 but 5 · h1, 2, 3i = h5, 10, 15i so the normal
vectors; thus also the planes; are not parallel. To find a common point we
need to find a common solution to

x + 2y + 3z = 4
5x + 6y + 7z = 8
one trick, which may work, is to set z = 0 this gives:

x + 2y = 4
5x + 6y = 8
In this case we are lucky and we find the solution x = −2, y = 3 so that
P = (−2, 3, 0). Check the answer by plugging into the equations!
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 9

1.3.2 Exercises
1. Compute the following dot products:

(a) h1, 2i · h3, 4i


(b) h1, 0, 1i · h2, 256, −2i
(c) h−12, −0.3, 2i · h1, 2, −4i

2. Compute the angles between the following pairs of vectors:

(a) h5, 6i and h−3, 2i


(b) h1, 2, 3i and h4, 5, 6i
(c) h−7, 8, −9, 2i and h1, −2, 0, 3i

3. Let ~v = h1, 1, 0i. Find two unit vectors ~u1 , ~u2 that are orthogonal to ~v .

4. (a) Find the equation of the plane passing trough P = (1, 1, 1) and
normal to ~n = (2, −1, −2). Express it in the form

z = ax + by + d.

Does this plane pass through the origin?


(b) Find the equation of a plane in R3 that is perpendicular to the
y-axis and passes through the point P = (−1, 3, −2). Does this
plane contain the points Q = (1, 1, 1) or R = (4, 3, 27 )?
(c) Find a normal vector for and a point on the plane

x + y + z = 1.

(d) Find a normal vector for and a point on the plane

3x − 2y + z = 13.

(e) Express as y = mx + b the line ` that passes through (−1, 1) with


normal vector h−1, 1i .
(f) What is the line ` in R2 passing through P = (2, 3) that is normal
to ~n = h−3, 0i? Can it be expressed as y = mx + b?

5. Consider the following picture


CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 10

y
C

D
B P

~n
A
x

Which of the following points A, B, C, D satisfy P~X · ~n = 0?

6. Using the properties of dot products. Show the following

~ ~a are orthogonal to ~v then so are α · w


(a) If the vectors w, ~ (α is a
scalar) and ~a + w.
~
(b) If Q is another point in a plane Π normal to ~n and containing P ,
then the equation of the plane containing Q and normal to ~n will
be identical.
(c) Find a formula involving the dot product to get the second entry
of any vector. For example the formula should produce 5 from
h3, 5, 2i.

7. Suppose you have goods such that the price per kilogram is given by a
vector P~ = hp1 , p2 , p3 i and let Q
~ = hq1 , q2 , q3 i be a vector of masses in
kilograms. What does P~ · Q ~ represent?
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 11

Answers:
1. (a) 11
(b) 0
−103
(c) 5
√ √ 
3
2. (a) arccos − 793 61 13 radians or approximately 96.11550356628541◦
√ √ 
16
(b) arccos 539 77 14 radians or approximately 12.933154491899106◦
√ √ 
17
(c) arccos − 924 22 14 radians or approximately 108.83768726421319◦

3. There are infinitely many good answers to this question, but h0, 0, 1i
works, or any unit vector ~u such that ~u · h1, 1, 0i = 0. Another choice
is √12 h−1, 1, 0i.

4. (a)

2x − y − 2z = −1
1 y
⇒z = + x −
2 2

The plane does not pass through the origin.


(b)
y=3
Does not contain Q, contains R.
(c)

−x+y =2
⇒y = x + 2

(d) ~n = h1, 1, 1i, P = (1, 0, 0).


(e) ~n = h3, −2, 1i, P = (0, 0, 13).
(f)
` : −3x = −6
That, is the vertical line x = 2. It cannot be expressed as y =
mx + b.
CHAPTER 1. MULTIVARIABLE FUNCTIONS; VECTORS 12

5. The points A, D, and P .

6. (a) If ~v · w
~ = 0 then

~v · (α · w)
~ = α (~v · w)
~ = α · 0 = 0.

If ~v · ~a = 0 as well, then

~v · (~a + w)
~ = ~v · ~a + ~v · w
~ = 0 + 0 = 0.

So both these other vectors are orthogonal to ~v .


(b) Let Q be another point on Π, then by definition ~n · P~Q = 0. In
~ = P~ + P~Q,
particular since Q
~ = ~n · (P~ + P~Q) = ~n · P~ + ~n · P~Q = ~n · P~ + 0,
~n · Q

we have the equivalence of formulas


~ = ~n · P~ ⇔ ~n · X
~n · X ~ = ~n · Q.
~

(c) The formula for the second entry of ~v is

~v · h0, 1, 0i .

7. The cost of ordering q1 kilos of the first good, q2 kilos of the second
good and q3 kilos of the third good.

Vous aimerez peut-être aussi