2.

3 Tangent Plane to a Surface 75

2.3 Tangent Plane to a Surface

∂f ∂f
In the previous section we mentioned that the partial derivatives ∂ x and ∂ y can be thought
of as the rate of change of a function z = f ( x, y) in the positive x and y directions, respectively.
dy
Recall that the derivative dx of a function y = f ( x) has a geometric meaning, namely as the
slope of the tangent line to the graph of f at the point ( x, f ( x)) in R2 . There is a similar
∂f ∂f
geometric meaning to the partial derivatives ∂ x and ∂ y of a function z = f ( x, y): given a
point (a, b) in the domain D of f ( x, y), the trace of the surface described by z = f ( x, y) in the
plane y = b is a curve in R3 through the point (a, b, f (a, b)), and the slope of the tangent line
∂f ∂f
L x to that curve at that point is ∂ x (a, b). Similarly, ∂ y (a, b) is the slope of the tangent line
L y to the trace of the surface z = f ( x, y) in the plane x = a (see Figure 2.3.1).

z z ∂f
z = f (x, y) slope = ∂y
(a, b)
∂f
(a, b, f (a, b)) slope = ∂x
(a, b) z = f (x, y)
(a, b, f (a, b))

Ly
Lx
b y y
0 0
a
(a, b) (a, b)
x x
D D
(a) Tangent line L x in the plane y = b (b) Tangent line L y in the plane x = a

Figure 2.3.1 Partial derivatives as slopes

dy
Since the derivative dx of a function y = f ( x) is used to find the tangent line to the graph
of f (which is a curve in R2 ), you might expect that partial derivatives can be used to define
a tangent plane to the graph of a surface z = f ( x, y). This indeed turns out to be the case.
First, we need a definition of a tangent plane. The intuitive idea is that a tangent plane “just
touches” a surface at a point. The formal definition mimics the intuitive notion of a tangent
line to a curve.

Definition 2.4. Let z = f ( x, y) be the equation of a surface S in R3 , and let P = (a, b, c) be

a point on S . Let T be a plane which contains the point P , and let Q = ( x, y, z) represent a
−−→
generic point on the surface S . If the (acute) angle between the vector PQ and the plane
T approaches zero as the point Q approaches P along the surface S , then we call T the
tangent plane to S at P .

Note that since two lines in R3 determine a plane, then the two tangent lines to the surface
z = f ( x, y) in the x and y directions described in Figure 2.3.1 are contained in the tangent
plane at that point, if the tangent plane exists at that point. The existence of those two
76 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

tangent lines does not by itself guarantee the existence of the tangent plane. It is possible
that if we take the trace of the surface in the plane x − y = 0 (which makes a 45◦ angle with
the positive x-axis), the resulting curve in that plane may have a tangent line which is not
in the plane determined by the other two tangent lines, or it may not have a tangent line
∂f ∂f
at all at that point. Luckily, it turns out4 that if ∂ x and ∂ y exist in a region around a point
(a, b) and are continuous at (a, b) then the tangent plane to the surface z = f ( x, y) will exist
at the point (a, b, f (a, b)). In this text, those conditions will always hold.
Suppose that we want an equation of the tangent plane T z
to the surface z = f ( x, y) at a point (a, b, f (a, b)). Let L x and z = f (x, y)
L y be the tangent lines to the traces of the surface in the
(a, b, f (a, b)) L
planes y = b and x = a, respectively (as in Figure 2.3.2), and y
suppose that the conditions for T to exist do hold. Then the Lx
equation for T is T y
0
A ( x − a) + B( y − b) + C ( z − f (a, b)) = 0 (2.4)
x
Figure 2.3.2 Tangent plane
where n = ( A, B, C ) is a normal vector to the plane T . Since
T contains the lines L x and L y , then all we need are vectors vx and v y that are parallel to L x
and L y , respectively, and then let n = vx × v y .
∂f ∂f z
Since the slope of L x is ∂ x (a, b), then the vector vx = (1, 0, ∂ x (a, b)) is ∂f
vx = (1, 0, ∂ x (a, b))
parallel to L x (since vx lies in the xz-plane and lies in a line with slope
∂f
∂x
(a,b) ∂f
1 =∂x
(a, b). See Figure 2.3.3). Similarly, the vector
∂f
∂f (a, b)
vy = (0, 1, ∂ y (a, b)) is parallel to L y . Hence, the vector ∂x
x
¯ ¯ 0
¯ i j k ¯ 1
¯ ¯
¯ ∂f ¯ ∂f ∂f
n = vx × v y = ¯ 1 0 ∂x
(a, b) ¯ = − ∂ x (a, b) i − ∂ y (a, b) j + k Figure 2.3.3
¯ ∂f ¯
¯ 0 1 (a, b) ¯
∂y

is normal to the plane T . Thus the equation of T is

∂f ∂f
− ∂ x (a, b) ( x − a) − ∂ y (a, b) ( y − b) + z − f (a, b) = 0 . (2.5)

Multiplying both sides by −1, we have the following result:

The equation of the tangent plane to the surface z = f ( x, y) at the point (a, b, f (a, b)) is
∂f ∂f
∂x
(a, b) ( x − a) + ∂ y (a, b) ( y − b) − z + f (a, b) = 0 (2.6)

4 See T AYLOR and M ANN, § 6.4.

 2.3 Tangent Plane to a Surface 77

Example 2.13. Find the equation of the tangent plane to the surface z = x2 + y2 at the point
(1, 2, 5).
∂f ∂f
Solution: For the function f ( x, y) = x2 + y2 , we have ∂x
= 2 x and ∂y
= 2 y, so the equation of
the tangent plane at the point (1, 2, 5) is

2(1)( x − 1) + 2(2)( y − 2) − z + 5 = 0 , or
2x + 4 y − z − 5 = 0 .

In a similar fashion, it can be shown that if a surface is defined implicitly by an equation

of the form F ( x, y, z) = 0, then the tangent plane to the surface at a point (a, b, c) is given by
the equation
∂F
∂x
(a, b, c) ( x − a) + ∂∂Fy (a, b, c) ( y − b) + ∂∂Fz (a, b, c) ( z − c) = 0 . (2.7)
Note that formula (2.6) is the special case of formula (2.7) where F ( x, y, z) = f ( x, y) − z.

Example 2.14. Find the equation of the tangent plane to the surface x2 + y2 + z2 = 9 at the
point (2, 2, −1).
∂F ∂F ∂F
Solution: For the function F ( x, y, z) = x2 + y2 + z2 − 9, we have ∂x
= 2 x, ∂y
= 2 y, and ∂z
= 2 z,
so the equation of the tangent plane at (2, 2, −1) is

2(2)( x − 2) + 2(2)( y − 2) + 2(−1)( z + 1) = 0 , or

2x + 2 y − z − 9 = 0 .

Exercises
A
For Exercises 1-6, find the equation of the tangent plane to the surface z = f ( x, y) at the
point P .

1. f ( x, y) = x2 + y3 , P = (1, 1, 2) 2. f ( x, y) = x y, P = (1, −1, −1)

3. f ( x, y) = x2 y, P = (−1, 1, 1) 4. f ( x, y) = xe y , P = (1, 0, 1)
p
5. f ( x, y) = x + 2 y, P = (2, 1, 4) 6. f ( x, y) = x2 + y2 , P = (3, 4, 5)

For Exercises 7-10, find the equation of the tangent plane to the given surface at the point
P.
³ p ´
2 y2 z2
7. x4 + 9 + 16 = 1, P = 1, 2, 2 311 8. x2 + y2 + z2 = 9, P = (0, 0, 3)
p
9. x2 + y2 − z2 = 0, P = (3, 4, 5) 10. x2 + y2 = 4, P = ( 3, 1, 0)
78 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

2.4 Directional Derivatives and the Gradient

∂f ∂f
For a function z = f ( x, y), we learned that the partial derivatives ∂ x and ∂ y represent the
(instantaneous) rate of change of f in the positive x and y directions, respectively. What
about other directions? It turns out that we can find the rate of change in any direction
using a more general type of derivative called a directional derivative.

Definition 2.5. Let f ( x, y) be a real-valued function with domain D in R2 , and let (a, b) be a
point in D . Let v be a unit vector in R2 . Then the directional derivative of f at (a, b) in
the direction of v, denoted by D v f (a, b), is defined as

f ((a, b) + hv) − f (a, b)

D v f (a, b) = lim (2.8)
h→0 h

Notice in the definition that we seem to be treating the point (a, b) as a vector, since we
are adding the vector hv to it. But this is just the usual idea of identifying vectors with their
terminal points, which the reader should be used to by now. If we were to write the vector v
as v = (v1 , v2 ), then
f (a + hv1 , b + hv2 ) − f (a, b)
D v f (a, b) = lim . (2.9)
h→0 h
∂f ∂f
From this we can immediately recognize that the partial derivatives ∂ x and ∂ y are special
cases of the directional derivative with v = i = (1, 0) and v = j = (0, 1), respectively. That is,
∂f ∂f
∂x
= D i f and ∂ y = D j f . Since there are many vectors with the same direction, we use a unit
vector in the definition, as that represents a “standard” vector for a given direction.
∂f ∂f
If f ( x, y) has continuous partial derivatives ∂ x and ∂ y (which will always be the case in
this text), then there is a simple formula for the directional derivative:

Theorem 2.2. Let f ( x, y) be a real-valued function with domain D in R2 such that the
∂f ∂f
partial derivatives ∂ x and ∂ y exist and are continuous in D . Let (a, b) be a point in D , and
let v = (v1 , v2 ) be a unit vector in R2 . Then

∂f ∂f
D v f (a, b) = v1 (a, b) + v2 (a, b) . (2.10)
∂x ∂y

∂f
Proof: Note that if v = i = (1, 0) then the above formula reduces to D v f (a, b) = ∂x
(a, b),
∂f
which we know is true since D i f = ∂x
, as we noted earlier. Similarly, for v = j = (0, 1) the
∂f ∂f
formula reduces to D v f (a, b) = ∂y
(a, b), which is true since D j f = ∂y
. So since i = (1, 0) and
2
j = (0, 1) are the only unit vectors in R with a zero component, then we need only show the
formula holds for unit vectors v = (v1 , v2 ) with v1 6= 0 and v2 6= 0. So fix such a vector v and
fix a number h 6= 0.
 2.4 Directional Derivatives and the Gradient 79

Then

f (a + hv1 , b + hv2 ) − f (a, b) = f (a + hv1 , b + hv2 ) − f (a + hv1 , b) + f (a + hv1 , b) − f (a, b) . (2.11)

Since h 6= 0 and v2 6= 0, then hv2 6= 0 and thus any number c between b and b + hv2 can be
written as c = b +α hv2 for some number 0 < α < 1. So since the function f (a+ hv1 , y) is a real-
valued function of y (since a + hv1 is a fixed number), then the Mean Value Theorem from
single-variable calculus can be applied to the function g( y) = f (a + hv1 , y) on the interval
[ b, b + hv2 ] (or [ b + hv2 , b] if one of h or v2 is negative) to find a number 0 < α < 1 such that

∂f g( b + hv2 ) − g( b) f (a + hv1 , b + hv2 ) − f (a + hv1 , b)

(a + hv1 , b + α hv2 ) = g ′ ( b + α hv2 ) = =
∂y b + hv2 − b hv2

and so
∂f
f (a + hv1 , b + hv2 ) − f (a + hv1 , b) = hv2 (a + hv1 , b + α hv2 ) .
∂y
By a similar argument, there exists a number 0 < β < 1 such that

∂f
f (a + hv1 , b) − f (a, b) = hv1 (a + β hv1 , b) .
∂x
Thus, by equation (2.11), we have
∂f ∂f
f (a + hv1 , b + hv2 ) − f (a, b) hv2 ∂ y (a + hv1 , b + α hv2 ) + hv1 ∂ x (a + β hv1 , b)
=
h h
∂f ∂f
= v2 (a + hv1 , b + α hv2 ) + v1 (a + β hv1 , b)
∂y ∂x

so by formula (2.9) we have

f (a + hv1 , b + hv2 ) − f (a, b)

D v f (a, b) = lim
h→0 h
· ¸
∂f ∂f
= lim v2 (a + hv1 , b + α hv2 ) + v1 (a + β hv1 , b)
h→0 ∂y ∂x

∂f ∂f ∂f ∂f
= v2 (a, b) + v1 (a, b) by the continuity of and , so
∂y ∂x ∂x ∂y

∂f ∂f
D v f (a, b) = v1 (a, b) + v2 (a, b)
∂x ∂y

after reversing the order of summation. QED

³ ´
∂f ∂f
Note that D v f (a, b) = v · ∂x
(a, b), ∂ y (a, b) . The second vector has a special name:
80 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

Definition 2.6. For a real-valued function f ( x, y), the gradient of f , denoted by ∇ f , is the
vector ³∂f ∂f ´
∇f = , (2.12)
∂x ∂ y
in R2 . For a real-valued function f ( x, y, z), the gradient is the vector
³∂f ∂f ∂f ´
∇f = , , (2.13)
∂x ∂ y ∂z

in R3 . The symbol ∇ is pronounced “del”.5

Corollary 2.3. D v f = v · ∇ f

2 3
´ directional derivative of f ( x, y) = x y + x y at the point (1, 2) in the
Example 2.15. ³Find the
direction of v = p1 , p1 .
2 2

Solution: We see that ∇ f = ( y2 + 3 x2 y, 2 x y + x3 ), so

³ ´
D v f (1, 2) = v · ∇ f (1, 2) = p1 , p1 · (22 + 3(1)2 (2), 2(1)(2) + 13 ) = 15
p
2 2 2

∂f ∂f
A real-valued function z = f ( x, y) whose partial derivatives ∂ x and ∂ y exist and are con-
tinuous is called continuously differentiable. Assume that f ( x, y) is such a function and that
∇ f 6= 0. Let c be a real number in the range of f and let v be a unit vector in R2 which is
tangent to the level curve f ( x, y) = c (see Figure 2.4.1).
y
v ∇f

f ( x, y) = c

x
0

Figure 2.4.1

5 Sometimes the notation grad( f ) is used instead of ∇ f .

 2.4 Directional Derivatives and the Gradient 81

The value of f ( x, y) is constant along a level curve, so since v is a tangent vector to this
curve, then the rate of change of f in the direction of v is 0, i.e. D v f = 0. But we know that
D v f = v · ∇ f = k v k k∇ f k cos θ , where θ is the angle between v and ∇ f . So since k v k = 1 then
D v f = k∇ f k cos θ . So since ∇ f 6= 0 then D v f = 0 ⇒ cos θ = 0 ⇒ θ = 90◦ . In other words, ∇ f ⊥ v,
which means that ∇ f is normal to the level curve.
In general, for any unit vector v in R2 , we still have D v f = k∇ f k cos θ , where θ is the angle
between v and ∇ f . At a fixed point ( x, y) the length k∇ f k is fixed, and the value of D v f then
varies as θ varies. The largest value that D v f can take is when cos θ = 1 (θ = 0◦ ), while the
smallest value occurs when cos θ = −1 (θ = 180◦ ). In other words, the value of the function
f increases the fastest in the direction of ∇ f (since θ = 0◦ in that case), and the value of
f decreases the fastest in the direction of −∇ f (since θ = 180◦ in that case). We have thus
proved the following theorem:

Theorem 2.4. Let f ( x, y) be a continuously differentiable real-valued function, with ∇ f 6= 0.

Then:

(a) The gradient ∇ f is normal to any level curve f ( x, y) = c.

(b) The value of f ( x, y) increases the fastest in the direction of ∇ f .

(c) The value of f ( x, y) decreases the fastest in the direction of −∇ f .

Example 2.16. In which direction does the function f ( x, y) = x y2 + x3 y increase the fastest
from the point (1, 2)? In which direction does it decrease the fastest?
Solution: Since ∇ f = (³y2 + 3 x´2 y, 2 x y + x3 ), then ∇ f (1, 2) = (10, 5) 6= 0. A unit vector
³ in´ that
∇f 2 p1 2 p1
direction is v = k∇ f k = p , . Thus, f increases the fastest in the direction of p , and
5 5 5 5
³ ´
−2 p
decreases the fastest in the direction of p , −1 .
5 5

Though we proved Theorem 2.4 for functions of two variables, a similar argument can
be used to show that it also applies to functions of three or more variables. Likewise, the
directional derivative in the three-dimensional case can also be defined by the formula D v f =
v· ∇f .

Example 2.17. The temperature T of a solid is given by the function T ( x, y, z) = e− x + e−2 y +

e4 z , where x, y, z are space coordinates relative to the center of the solid. In which direction
from the point (1, 1, 1) will the temperature decrease the fastest?
Solution: Since ∇ f = (− e− x , −2 e−2 y , 4 e4 z ), then the temperature will decrease the fastest in
the direction of −∇ f (1, 1, 1) = ( e−1 , 2 e−2 , −4 e4 ).
82 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

Exercises
A
For Exercises 1-10, compute the gradient ∇ f .
1
1. f ( x, y) = x2 + y2 − 1 2. f ( x, y) =
x 2 + y2
p
3. f ( x, y) = x 2 + y2 + 4 4. f ( x, y) = x2 e y

5. f ( x, y) = ln( x y) 6. f ( x, y) = 2 x + 5 y

7. f ( x, y, z) = sin( x yz) 8. f ( x, y, z) = x2 e yz
p
9. f ( x, y, z) = x2 + y2 + z2 10. f ( x, y, z) = x2 + y2 + z2

For ³Exercises
´ 11-14, find the directional derivative of f at the point P in the direction of
v= p1 , p1 .
2 2
1
11. f ( x, y) = x2 + y2 − 1, P = (1, 1) 12. f ( x, y) = , P = (1, 1)
x 2 + y2
p
13. f ( x, y) = x2 + y2 + 4, P = (1, 1) 14. f ( x, y) = x2 e y , P = (1, 1)

For ³Exercises 15-16,

´ find the directional derivative of f at the point P in the direction of
1 1 1
v= p ,p ,p .
3 3 3
15. f ( x, y, z) = sin( x yz), P = (1, 1, 1) 16. f ( x, y, z) = x2 e yz , P = (1, 1, 1)

17. Repeat Example 2.16 at the point (2, 3).

18. Repeat Example 2.17 at the point (3, 1, 2).

B
For Exercises 19-26, let f ( x, y) and g( x, y) be continuously differentiable real-valued func-
tions, let c be a constant, and let v be a unit vector in R2 . Show that:

19. ∇( c f ) = c ∇ f 20. ∇( f + g) = ∇ f + ∇ g
g ∇f − f ∇g
21. ∇( f g) = f ∇ g + g ∇ f 22. ∇( f / g) = if g( x, y) 6= 0
g2
23. D −v f = −D v f 24. D v ( c f ) = c D v f

25. D v ( f + g) = D v f + D v g 26. D v ( f g) = f D v g + g D v f
p
27. The function r ( x, y) = x2 + y2 is the length of the position vector r = x i + y j for each
1
point ( x, y) in R2 . Show that ∇ r = r when ( x, y) 6= (0, 0), and that ∇( r 2 ) = 2 r.
r
 2.5 Maxima and Minima 83

2.5 Maxima and Minima

The gradient can be used to find extreme points of real-valued functions of several variables,
that is, points where the function has a local maximum or local minimum. We will consider
only functions of two variables; functions of three or more variables require methods using
linear algebra.

Definition 2.7. Let f ( x, y) be a real-valued function, and let (a, b) be a point in the domain
of f . We say that f has a local maximum at (a, b) if f ( x, y) ≤ f (a, b) for all ( x, y) inside some
disk of positive radius centered at (a, b), i.e. there is some sufficiently small r > 0 such that
f ( x, y) ≤ f (a, b) for all ( x, y) for which ( x − a)2 + ( y − b)2 < r 2 .
Likewise, we say that f has a local minimum at (a, b) if f ( x, y) ≥ f (a, b) for all ( x, y)
inside some disk of positive radius centered at (a, b).
If f ( x, y) ≤ f (a, b) for all ( x, y) in the domain of f , then f has a global maximum at
(a, b). If f ( x, y) ≥ f (a, b) for all ( x, y) in the domain of f , then f has a global minimum at
(a, b).

Suppose that (a, b) is a local maximum point for f ( x, y), and that the first-order partial
derivatives of f exist at (a, b). We know that f (a, b) is the largest value of f ( x, y) as ( x, y)
goes in all directions from the point (a, b), in some sufficiently small disk centered at (a, b).
In particular, f (a, b) is the largest value of f in the x direction (around the point (a, b)), that
is, the single-variable function g( x) = f ( x, b) has a local maximum at x = a. So we know that
∂f ∂f
g ′ (a) = 0. Since g ′ ( x) = ∂ x ( x, b), then ∂ x (a, b) = 0. Similarly, f (a, b) is the largest value of f
∂f
near (a, b) in the y direction and so ∂y
(a, b) = 0. We thus have the following theorem:

∂f ∂f
Theorem 2.5. Let f ( x, y) be a real-valued function such that both ∂ x (a, b) and ∂ y (a, b) exist.
Then a necessary condition for f ( x, y) to have a local maximum or minimum at (a, b) is that
∇ f (a, b) = 0.

Note: Theorem 2.5 can be extended to apply to functions of three or more variables.
A point (a, b) where ∇ f (a, b) = 0 is called a critical point for the function f ( x, y). So given
∂f
a function f ( x, y), to find the critical points of f you have to solve the equations ∂ x ( x, y) = 0
∂f
and ∂ y ( x, y) = 0 simultaneously for ( x, y). Similar to the single-variable case, the necessary
condition that ∇ f (a, b) = 0 is not always sufficient to guarantee that a critical point is a local
maximum or minimum.

∂f
Example 2.18. The function f ( x, y) = x y has a critical point at (0, 0): ∂x
= y = 0 ⇒ y = 0, and
∂f
∂y
= x = 0 ⇒ x = 0, so (0, 0) is the only critical point. But clearly f does not have a local
maximum or minimum at (0, 0) since any disk around (0, 0) contains points ( x, y) where the
values of x and y have the same sign (so that f ( x, y) = x y > 0 = f (0, 0)) and different signs (so
that f ( x, y) = x y < 0 = f (0, 0)). In fact, along the path y = x in R2 , f ( x, y) = x2 , which has a
84 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

local minimum at (0, 0), while along the path y = − x we have f ( x, y) = − x2 , which has a local
maximum at (0, 0). So (0, 0) is an example of a saddle point, i.e. it is a local maximum in one
direction and a local minimum in another direction. The graph of f ( x, y) is shown in Figure
2.5.1, which is a hyperbolic paraboloid.

100

50

0
z

-50

-10
-100-10 -5
-5 0 x
0 5 5
y 10 10

Figure 2.5.1 f (x, y) = x y, saddle point at (0, 0)

The following theorem gives sufficient conditions for a critical point to be a local maximum
or minimum of a smooth function (i.e. a function whose partial derivatives of all orders exist
and are continuous), which we will not prove here.6

Theorem 2.6. Let f ( x, y) be a smooth real-valued function, with a critical point at (a, b) (i.e.
∇ f (a, b) = 0). Define
∂2 f ∂2 f ³ ∂2 f ´2
D = 2 (a, b) 2 (a, b) − (a, b)
∂x ∂y ∂ y ∂x
Then
∂2 f
(a) if D > 0 and ∂ x2
(a, b) > 0, then f has a local minimum at (a, b)

∂2 f
(b) if D > 0 and ∂ x2
(a, b) < 0, then f has a local maximum at (a, b)

(c) if D < 0, then f has neither a local minimum nor a local maximum at (a, b)

(d) if D = 0, then the test fails.

6 See T AYLOR and M ANN, § 7.6.

 2.5 Maxima and Minima 85

If condition (c) holds, then (a, b) is a saddle point. Note that the assumption that f ( x, y) is
smooth means that ¯ 2 ¯
¯ ∂ f ∂2 f ¯
¯
¯ ∂ x2 ( a, b ) ( a, b ) ¯
¯ ∂ y ∂x ¯
¯
D=¯ ¯
¯ ∂2 f ∂2
f ¯
¯ (a, b) (a, b) ¯¯
¯ 2
∂x ∂ y ∂y
³ 2 ´2
∂2 f ∂2 f ∂2 f ∂2 f ∂ f ∂2 f
since ∂ y ∂ x = ∂ x ∂ y . Also, if D > 0 then ∂ x2 (a, b) ∂ y2 (a, b) = D + ∂ y ∂ x (a, b) > 0, and so ∂ x2 (a, b)
∂2 f
and ∂ y2
(a, b) have the same sign. This means that in parts (a) and (b) of the theorem one
∂2 f ∂2 f
can replace ∂ x2 (a, b) by ∂ y2 (a, b) if desired.

Example 2.19. Find all local maxima and minima of f ( x, y) = x2 + x y + y2 − 3 x.

Solution: First find the critical points, i.e. where ∇ f = 0. Since

∂f ∂f
= 2x + y − 3 and = x + 2y
∂x ∂y

then the critical points ( x, y) are the common solutions of the equations

2x + y − 3 = 0
x + 2y =0

which has the unique solution ( x, y) = (2, −1). So (2, −1) is the only critical point.
To use Theorem 2.6, we need the second-order partial derivatives:

∂2 f ∂2 f ∂2 f
=2 , =2 , =1
∂ x2 ∂ y2 ∂ y ∂x

and so
∂2 f ∂2 f ³ ∂2 f ´2
D = 2
(2, −1) 2
(2, −1) − (2, −1) = (2)(2) − 12 = 3 > 0
∂x ∂y ∂ y ∂x
∂2 f
and ∂ x2
(2, −1) = 2 > 0. Thus, (2, −1) is a local minimum.

Example 2.20. Find all local maxima and minima of f ( x, y) = x y − x3 − y2 .

Solution: First find the critical points, i.e. where ∇ f = 0. Since

∂f ∂f
= y − 3 x2 and = x − 2y
∂x ∂y
86 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

then the critical points ( x, y) are the common solutions of the equations

y − 3 x2 = 0
x − 2y = 0

The first equation yields y = 3 x2 , substituting that into the second equation yields x −6 x2 = 0,
¡ ¢2 1
which has the solutions x = 0 and x = 16 . So x = 0 ⇒ y = 3(0) = 0 and x = 61 ⇒ y = 3 61 = 12 .
¡1 1 ¢
So the critical points are ( x, y) = (0, 0) and ( x, y) = 6 , 12 .
To use Theorem 2.6, we need the second-order partial derivatives:

∂2 f ∂2 f ∂2 f
= −6 x , = −2 , =1
∂ x2 ∂ y2 ∂ y ∂x

So
∂2 f ∂2 f ³ ∂2 f ´2
D = 2
(0, 0) 2
(0, 0) − (0, 0) = (−6(0))(−2) − 12 = −1 < 0
∂x ∂y ∂ y ∂x
and thus (0, 0) is a saddle point. Also,

∂2 f ¡ 1 ¢ ∂2 f ¡ 1 ¢ ³ ∂2 f ¡ 1 1 ¢´2 ¡1¢
D = 1
6 , 12
1
6 12 − ∂ y ∂ x 6 , 12
, = (−6 )(−2) − 12 = 1 > 0
∂ x2 ∂ y2 6

∂2 f ¡ 1 1 ¢ ¡1 1
¢
and ,
∂ x2 6 12
= −1 < 0. Thus, 6 , 12 is a local maximum.

Example 2.21. Find all local maxima and minima of f ( x, y) = ( x − 2)4 + ( x − 2 y)2 .
Solution: First find the critical points, i.e. where ∇ f = 0. Since

∂f ∂f
= 4( x − 2)3 + 2( x − 2 y) and = −4( x − 2 y)
∂x ∂y

then the critical points ( x, y) are the common solutions of the equations

4( x − 2)3 + 2( x − 2 y) = 0
−4( x − 2 y) = 0

The second equation yields x = 2 y, substituting that into the first equation yields 4(2 y − 2)3 =
0, which has the solution y = 1, and so x = 2(1) = 2. Thus, (2, 1) is the only critical point.
To use Theorem 2.6, we need the second-order partial derivatives:

∂2 f ∂2 f ∂2 f
= 12( x − 2)2 + 2 , =8 , = −4
∂ x2 ∂ y2 ∂ y ∂x
 2.5 Maxima and Minima 87

So
∂2 f ∂2 f ³ ∂2 f ´2
D = 2
(2, 1) 2
(2, 1) − (2, 1) = (2)(8) − (−4)2 = 0
∂x ∂y ∂ y ∂x
and so the test fails. What can be done in this situation? Sometimes it is possible to examine
the function to see directly the nature of a critical point. In our case, we see that f ( x, y) ≥ 0
for all ( x, y), since f ( x, y) is the sum of fourth and second powers of numbers and hence must
be nonnegative. But we also see that f (2, 1) = 0. Thus f ( x, y) ≥ 0 = f (2, 1) for all ( x, y), and
hence (2, 1) is in fact a global minimum for f .

2
+ y2 )
Example 2.22. Find all local maxima and minima of f ( x, y) = ( x2 + y2 ) e−( x .
Solution: First find the critical points, i.e. where ∇ f = 0. Since

∂f 2
+ y2 )
= 2 x(1 − ( x2 + y2 )) e−( x
∂x
∂f 2
+ y2 )
= 2 y(1 − ( x2 + y2 )) e−( x
∂y

then the critical points are (0, 0) and all points ( x, y) on the unit circle x2 + y2 = 1.
To use Theorem 2.6, we need the second-order partial derivatives:

∂2 f 2
+ y2 )
= 2[1 − ( x2 + y2 ) − 2 x2 − 2 x2 (1 − ( x2 + y2 ))] e−( x
∂ x2
∂2 f 2
+ y2 )
= 2[1 − ( x2 + y2 ) − 2 y2 − 2 y2 (1 − ( x2 + y2 ))] e−( x
∂ y2

∂2 f 2
+ y2 )
= −4 x y[2 − ( x2 + y2 )] e−( x
∂ y ∂x

∂2 f
At (0, 0), we have D = 4 > 0 and ∂ x2 (0, 0) = 2 > 0, so (0, 0) is a local minimum. However, for
points ( x, y) on the unit circle x2 + y2 = 1, we have

D = (−4 x2 e−1 )(−4 y2 e−1 ) − (−4 x ye−1 )2 = 0

and so the test fails. If we look at the graph of f ( x, y), as shown in Figure 2.5.2, it looks like
we might have a local maximum for ( x, y) on the unit circle x2 + y2 = 1. If we switch to using
polar coordinates ( r, θ ) instead of ( x, y) in R2 , where r 2 = x2 + y2 , then we see that we can write
2 2
f ( x, y) as a function g( r ) of the variable r alone: g( r ) = r 2 e−r . Then g ′ ( r ) = 2 r (1 − r 2 ) e−r ,
so it has a critical point at r = 1, and we can check that g ′′ (1) = −4 e−1 < 0, so the Second
Derivative Test from single-variable calculus says that r = 1 is a local maximum. But r = 1
corresponds to the unit circle x2 + y2 = 1. Thus, the points ( x, y) on the unit circle x2 + y2 = 1
are local maximum points for f .
88 CHAPTER 2. FUNCTIONS OF SEVERAL VARIABLES

0.4
0.35
0.3
0.25
z 0.2
0.15 -3
0.1 -2
0.05
0 -1
-3
-2 0
-1 x
1
0
1 2
y 2
3 3
2 + y2 )
Figure 2.5.2 f (x, y) = (x2 + y2 )e−( x

Exercises
A
For Exercises 1-10, find all local maxima and minima of the function f ( x, y).

1. f ( x, y) = x3 − 3 x + y2 2. f ( x, y) = x3 − 12 x + y2 + 8 y

3. f ( x, y) = x3 − 3 x + y3 − 3 y 4. f ( x, y) = x3 + 3 x2 + y3 − 3 y2

5. f ( x, y) = 2 x3 + 6 x y + 3 y2 6. f ( x, y) = 2 x3 − 6 x y + y2
p
7. f ( x, y) = x2 + y2 8. f ( x, y) = x + 2 y

9. f ( x, y) = 4 x2 − 4 x y + 2 y2 + 10 x − 6 y 10. f ( x, y) = −4 x2 + 4 x y − 2 y2 + 16 x − 12 y

B
11. For a rectangular solid of volume 1000 cubic meters, find the dimensions that will min-
imize the surface area. (Hint: Use the volume condition to write the surface area as a
function of just two variables.)

12. Prove that if (a, b) is a local maximum or local minimum point for a smooth function
f ( x, y), then the tangent plane to the surface z = f ( x, y) at the point (a, b, f (a, b)) is parallel
to the x y-plane. (Hint: Use Theorem 2.5.)

C
13. Find three positive numbers x, y, z whose sum is 10 such that x2 y2 z is a maximum.