CNQ#10 PDF

10
Multivariable Calculus*
*This chapter is based on Calculus for Life Sciences, by Rodolfo De Sapio, and
Calculus by H. Anton et al.
So far we have only considered functions of one independent variable -

for instance, those of the form f (x) or y(t ). In this chapter we extend this to
functions of several independent variables.
10.1 Functions of two or more independent variables
An example of a function of more than one variable that arises in the life
sciences is the relation that exists between an animal’s body’s surface area,
weight and height. Such a relation is used by physiologists in metabolism
studies. We examine this instance by assigning particular values:
• let w denote weight measured in kilograms,

• let h be height measured in centimetres, and
• let s denote body surface area measured in square meters.
For humans, one function that scientists have proposed in this case is ex-
pressed by the equation
s = (0.007184)w0.425 h0.725 . (10.1)
Note that this equation is used to calculate body surface area from a person’s
height and weight. We say that the right hand side of the equation defines a
function f (w, h) of two variables w and h, with the function rule
f (w, h) = (0.007184)w0.425 h0.725 .
The relationship s = f (w, h) is accurate for values of weight w from 20 kg

to 110 kg and for values of height h from 100 cm to 200 cm.
Example 10.1 An eight year old child is h = 130 cm tall and weighs w =
35 kg. According to Eqn. 10.1, what is the surface area of this child?
290 I N T E G R A L C A L C U L U S W I T H A P P L I C AT I O N S F O R L I F E S C I E N C E S
Solution. We verify that the given height is within the domain of valid-
ity: 100 ≤ h = 130 ≤ 200 cm; similarly, 20 ≤ 35 = w ≤ 110. We substitute in
and find:
s = f (35, 130) = (0.007184)(350.425 )(1300.725 ) = 1.1097 m2
Example 10.2 A given horse is 1.5 m tall and weighs 600 kg. Find the
surface area of this horse.
Solution. While the height of the horse is within the range of valid values
(100 ≤ 150 ≤ 200), the weight is certainly outside. Thus, we cannot
use Eqn. 10.1 to determine the surface area of this horse.
With this example in mind, we can now formulate a general notion of a
function of two independent variables. Suppose the equation
z = f (x, y)
defines z as a function f of the independent variables x and y if, for each

point (x, y) in some fixed region D in the plane, f (x, y) is well-defined.
The notation f (x, y) is a mathematical expression with variables x and y: it
is a rule that associates precisely one value f (x, y) of z with each pair of num-
bers (x, y) in the region D. The variables x and y are called the independent
variables and the variable z is the dependent variable. The region D is called
the domain of the function f .
Example 10.3 What is the domain D of the surface area function described h
above in Eqn. 10.1? Plot this domain. 200
Solution. The domain D of this function is all points (w, h) of validity, i.e. 150
those such that 20 ≤ w ≤ 110 and 100 ≤ h ≤ 200. This region is depicted in
100
Figure 10.1.
50
Example 10.4 Let D = {(x, y) : 0 ≤ x ≤ 1, 0 ≤ y ≤ 1} be the domain for the
w
function 20 40 60 80 100 120
f (x, y) = x + y. Figure 10.1: Domain D = {(w, h) : 20 ≤ w ≤

120, 100 ≤ h ≤ 200}.
Graph the domain D of this function. Can you determine its range?
Solution. The domain D is pictured in Figure 10.2. To find the range

of f , we need to determine the values f can take on when we plug in all y
points (x, y) from the domain D. We can observe that f (x, y) is smallest when
we plug in (x, y) = (0, 0), and largest when we plug in (x, y) = (1, 1) (this
is straightforward due to the simple nature of the function described). We 1
note that f (0, 0) = 0 and f (1, 1) = 2, so the range of f is all values z such
that 0 ≤ z ≤ 2. .
x
1
Figure 10.2: Domain D = {(x, y) : 0 ≤ x ≤
1, 0 ≤ y ≤ 1}.
M U LT I VA R I A B L E C A L C U L U S 291
Notice in Example 10.4 that the result f (x, y) = 1 can be achieved in many
different ways:
f (1, 0) = 1, f (0, 1) = 1, f (0.5, 0.5) = 1, f (0.2, 0.8) = 1,
and so on. A desire to describe such a set of points (x, y) in D such that f (x, y) =
c for some constant c motivates our next section.
Contour lines
Any function z = f (x, y) is determined by a complete table of values - one
value z for each point (x, y) in the domain D. For Eqn. 10.1, the domain
Weight Height Surface area
was depicted by the rectangle in Figure 10.1. Some numerical values as- w kg h cm s m2
sociated with function 10.1, are given in Table 10.1. The curves shown in 40 120 1.108
50 160 1.500
Figure 10.3 are called contour lines (which in general can be straight or
50 170 1.569
curved). Each contour line is a curve along which the value s = f (w, h) 55 180 1.700
is the same, i.e. a constant. For example, the contour line that passes 75 150 1.700
80 140 1.664
through the points (55, 180) and (75, 150) consists of all points (w, h) such 100 180 2.195
that f (w, h) = 1.7 m2 .
Table 10.1: Associated values of weight,
height and surface area in humans.
h (cm)
200
f (w, h) = 1.7
180
Figure 10.3: Contour lines of the func-
tion f (w, h) = (0.007184)w0.425 h0.725 that
160 gives the relation s = f (w, h) between body
surface area s (measured in square meters),
weight w, and height h of a human. Such a
140 chart is used by physiologists in metabolism

studies. Data are adapted from E.F. Dubois,
Square meters as indicated in Table 10.1.
120
0.8 1.1
w (kg)
40 60 80 100
Example 10.5 We return to Example 10.4. Let D = {(x, y) : 0 ≤ x ≤ 1, 0 ≤

y ≤ 1} be the domain for the function
f (x, y) = x + y.
Plot the contour line such that f (x, y) = 1.

y
1.2
1
0.8
0.6
Solution. In our exposition, we already saw that the points (1, 0), (0, 1), (0.5, 0.5) 0.4
and (0.2, 0.8) all satisfy f (x, y) = 1. Our contour line contains all of these 0.2
points, and those such that x
0.2 0.4 0.6 0.8 1 1.2
Figure 10.4: Contour line f (x, y) = x + y = 1.
f (x, y) = x + y = 1 ⇒ y = 1 − x.
The domain limits both our x and y values to [0, 1]. The resulting contour line
is plotted in Figure 10.4.
In general, given a function z = f (x, y) defined in a region D and a con-
stant c, then the set of all points (x, y) in D that satisfy the equation f (x, y) = c
is called a contour line or level curve. Thus, a contour line is just a curve (x, y) f (x, y) = x2 + y2
in the region D – a property we can verify in each of the examples above. (0, 0) 0
Indeed, the contour line f (w, h) = 1.5 in Figure 10.3 is the curve passing (1, 0) 1
(2, 1) 5
through the point (50, 160), where the function f (w, h) is given by the right (−1, 1) 2
hand side of Eqn. 10.1. (0, −1) 1
(4, −2) 20
(−1, −1) 2
Example 10.6 Consider the function z = x2 + y2 with the function rule f (x, y) =
x2 + y2 defined for all points (x, y) in the plane. For f (x, y), Table 10.2: Values of the function f (x, y) =
x 2 + y2 .
(a) find some select values of this function;
y
(b) plot several contour lines in the xy-plane.
f (x, y) = 9
Solution. 2
(a) This function is defined for all points in the xy-plane, so we can simply f (x, y) = 1
substitute in any values for x and‘y and get the corresponding f (x, y) value. x
−2 2
Some of these are tabulated in Table 10.2.
(b) For any positive constant c, the contour line f (x, y) = x2 + y2 = c is the −2 f (x, y) = 4
√
circle with centre (0, 0) and radius c. In particular, for c = 4, the contour
line f (x, y) = 4 is the circle x2 + y2 = 4 of radius 2. Additional contour
Figure 10.5: The contour lines of the
lines are shown in Figure 10.5. Notice that the contour line f (x, y) = 0 function f (x, y) = x2 + y2 are concentric
is a single point, the origin (0, 0). There are no contour lines f (x, y) = c circles with centres at the origin.
for c < 0 because f (x, y) = x2 + y2 ≥ 0 for all point (x, y).
(x, y, f (x, y))
Graphs
Functions of one independent variable are graphed in the xy-plane deter- f (x, y)
mined by the x- and y-axes. y
D
The graph of a function y = f (x) is a set of points (x, f (x)), where x (x, y)
is a possible value of an independent variable in the domain of f (x).

x
The graph of f (x) may form a curve or it can be represented by a set of Figure 10.6: The graph of a function f is the
discrete points in the xy-plane. surface in space consisting of all points of
the form (x, y, f (x, y)) where (x, y) is a point
Once we start considering functions of two variables, their domains form in the domain D of f .
regions in the xy-plane. A third axis is needed to create a 3-dimensional
xyz-space. For every point (x0 , y0 ) in the domain D of a function z =

f (x, y), we calculate the value z0 = f (x0 , y0 ) in the range of the function.
Then the graph of the function f is defined to be the set of points of the
form (x, y, f (x, y)) in xyz-space, where (x, y) is a point in the region D.
The graph may form a surface. Special software may be used to draw
some complicated surfaces. The graph of a general function f is shown in
Figure 10.6.
Figure 10.7: The graph of the func-

tion f (x, y) = x2 + y2 .
c
√
c
x f (x, y) = c y
The graph of the function f (x, y) = x2 + y2 in Example 10.6 is shown in

Figure 10.7; notice that the intersection of the graph with the plane z = c,
√
where c denotes a positive constant, is a circle of radius c that is c units
above the contour line x2 + y2 = c in the xy-plane.
In general, a contour line (level curve) f (x, y) = c of a function f may be
obtained from the graph of f by intersecting the graph with the plane z = c, as
shown in Figure 10.8.
z y
z=c
Figure 10.8: In (a), the intersection of the
graph z = f (x, y) with the plane z = c
y (where c is a positive constant) is a curve in
space directly above the line f (x, y) = c in
the xy-plane. As the plane is moved up and
down, we obtain a family of contour lines,
c
as shown in (b).
c
f (x, y) = c
x
(a) x (b)
Level curves have wide applications in various fields. One common

example of level curves occurs in topographic maps of mountainous regions,

such as the map in Figure 10.9.
Figure 10.9: A topographical map with level

curves (open source TopoWeb). Path A has
steep parts C and D, and Path B - has steep
part E.
The level curves are curves of constant elevation above sea level. You
would never ascend or descend if you walk along the paths that correspond
to the contour lines. Other common examples are isothermals–level curves
that join locations with the same temperature and isobars–lines of constant
pressure.
In Math 154, we recall the graph of ax + by = c (where a and b are not both
zero) is a straight line. This result is also extended to the 3-dimensional case.
The graph of
ax + by + cz = d
is a plane if a, b, and c are not all zero.
Example 10.7 Graph 2x + y + z = 8.
Solution. The graph of this equation is a plane. When graphing lines, we of-
ten used x- and y- intercepts. A similar approach can be used when graphing
a plane. To find the x-intercept, which is the point where the graph crosses the
x-axis, let y = 0 and z = 0. Then 2x + 0 + 0 = 8, x = 4. The x-intercept is the
point (4, 0, 0). Letting x = 0 and y = 0 gives the y-intercept (0, 8, 0), while
x = 0 and y = 0 lead to the z-intercept (0, 0, 8). The plane through these three
points includes the triangular surface shown in Figure 10.10.
The plane does not stop at the axes but extends without bound. Another
way to represent the equation of this plane would be f (x, y) = 8 − 2x − y.
Any contour line f (x, y) = c of the function in Example 10.7 is a straight
line with equation 2x + y = c or, solving for y, y = c − 2x. These contour lines
are shown in Figure 10.11.
z Figure 10.10: The triangular portion of the

plane 2x + y + z = 8.
(0, 0, 8)
2x + y + z = 8
(4, 0, 0) (0, 8, 0) y
x
y
Example 10.8 Sketch the graph of the function f (x, y) = 3 − y2 , and then find
c
the contour lines of f (x, y) = 3 − y2 .
y = c − 2x
Solution. First, notice that this function does not depend on the variable x!
The intersection of each plane x = x0 with the surface z = 3 − y2 is a parabola
opening downward, as shown in Figure 10.12. The equation of each parabola
is z = 3 − y2 . Let us graph the parabola in the yz-plane (i.e. when x = 0). x
This parabola intersects the z-axis (when y = 0) at z = 3 and opens
y = −2x
downward. The surface is obtained by moving this parabola in the direction
Figure 10.11: Contour lines 2x + y = c of the
parallel to the x-axis as shown in Figure 10.12. function f (x, y) = 8 − 2x − y, whose graph is
We next determine the contour lines of f (x, y) = 3 − y2 . If c is a constant, shown in Figure 10.10. Each contour line is
√ a straight line with equation y = c − 2x.
then the contour line 3 − y2 = c consists of the two lines y = ± 3 − c parallel
to the x-axis, provided that c < 3. The contour line 3 − y2 = 3 is the x-axis
(where y = 0).
z Figure 10.12: The graph of f (x, y) = 3 − y2

z
x = x0
y
z = 3 − y2
y
x
The next example concerns a function with elliptic contour lines.

Example 10.9 Sketch the contour lines f (x, y) = 3, f (x, y) = 4, and
f (x, y) = 5 of the function
f (x, y) = x2 + 4y2 .
y
Solution. For any positive number c, the contour line x2 + 4y2 = c is an ellipse
with centre at the origin, as shown in Figure 10.13. This ellipse intersects
√ √
the x-axis (when y = 0) in the two point ( c, 0) and (− c, 0); it intersects 2
√ √ √
the y-axis (when x = 0) in the two points (0, c/2) and (0, − c/2). The 3 x
ellipses for the choices c = 3, 4, and 5 are shown in Figure 10.13. 0
√
Note: in this chapter’s introduction we used the term “functions of several 5
independent variables,” and although the word “several” often refers to
an indefinite number greater than two, we shall restrict our attention in
Figure 10.13: The contour lines x2 + 4y2 = c
this chapter to the study of functions of two independent variables. It is are ellipses with centre at the origin. The
conventional for mathematicians to use “several” even for two variables. contour lines for c = 3, 4, 5 are shown.
The fundamental ideas that distinguish the calculus of several variables
from the calculus of just one variable are clearly apparent in the study of
functions of just two independent variables. The ideas and results that hold in
studying two variables can be generalized naturally to the study of functions
of three or more independent variables.
10.2 Partial derivatives and applications
In order to introduce partial derivatives, let us first suppose that z = f (x, y) is a

function of two variables and (x0 , y0 ) is a point in its domain. Our goal is to
study the rate of change of f (x, y) at (x0 , y0 ) in the x-direction by choosing a
change in x: a change from x0 to x0 + Δx. We keep y fixed at y0 . Notice: the
point (x0 + Δx, y0 ) is near (x0 , y0 ) and the corresponding change in the value
of the function is,
Δz = f (x0 + Δx, y0 ) − f (x0 , y0 ).
The average rate of change of f (x, y) as (x, y) moves parallel to the x-axis
from (x0 , y0 ) to (x0 + Δx, y0 ) is defined by:
Δz f (x0 + Δx, y0 ) − f (x0 , y0 )

= .
Δx Δx
This change is depicted in Figure 10.14.
Definition 10.1 The instantaneous rate of change of f (x, y) in the x-direction

at (x0 , y0 ) is defined to be the limit of this difference quotient as Δx ap-
proaches zero:
Δz f (x0 + Δx, y0 ) − f (x0 , y0 )

lim = lim . (10.2)
Δx→0 Δx Δx→0 Δx
This limit is called the partial derivative of f (x, y) with respect to x at (x0 , y0 )
and is denoted by

∂ f f (x0 + Δx, y0 ) − f (x0 , y0 )
= lim .
∂ x (x0 ,y0 ) Δx→0 Δx
y Figure 10.14: As a point (x, y) moves

parallel to the x-axis from (x0 , y0 ) to (x0 +
Δx, y0 ), the average rate of change of the
value f (x, y) of the function f is
f (x0 + Δx, y0 ) − f (x0 , y0 )
(x0 , y0 ) (x0 + Δx, y0 ) .
y0 Δx
The limit of this average rate of change,
as Δx approaches zero, is the partial
derivative ∂∂ xf at (x0 , y0 ).
Δx
x
x0 x0 + Δx
As is typical, other notations may be used for Eqn. 10.2. These include:

∂ z
and fx (x0 , y0 ).
∂ x (x0 ,y0 )
Note: the symbol ∂ in the partial derivative ∂ f /∂ x is a rounded “d” and we

read the expression we have just written as “partial of f with respect to x” or
“partial d f by dx”. Note that an alternative way is saying “del f del x”.
When the point (x0 , y0 ) varies throughout the domain f (x, y), we get a
new function of x and y and denote it simply by ∂ f /∂ x and call it the partial
derivative of f (x, y) with respect to x. It is given by
∂f f (x + Δx, y) − f (x, y)
= lim .
∂ x Δx→0 Δx
Example 10.10 Let f (x, y) = x2 + x3 y2 . Find its partial derivative with

respect to x at the point (x0 , y0 ).
Solution. Referring to the definition of (∂ f /∂ x)(x0 ,y0 ) , we note that we are

actually taking the derivative of the function f (x, y0 ) of x alone because we
are keeping y fixed at the value y0 . This means that all our previous rules for
functions of one variable apply!
Thus, the derivative of f (x, y0 ) = x2 + x3 y20 with respect to x is 2x + 3x2 y20 .
In order to get the partial derivative of f (x, y) with respect to x at (x0 , y0 )
we simply substitute x0 for x in the expression 2x + 3x2 y2 to get

∂ f
= 2x0 + 3x02 y20 ,
∂ x (x0 ,y0 )
as required.
Note: we of course can find the partial derivative of f (x, y) (described

in Example 10.10) with respect to x at a particular point. In particular,
if (x0 , y0 ) = (2, 3), then

∂ f
= 2 · 2 + 3 · 22 · 32 = 112.
∂ x (2,3)
Example 10.11 Find the partial derivative with respect to x of f (x, y) =

x2 sin y + y2 ln x.
Solution. Since we are not asked to find the partial derivative of f (x, y) at any
particular point, we compute the derivative of
f (x, y) = x2 sin y + y2 ln x
with respect to x by treating y as a constant in our calculations. Thus, we get:

∂f 1
= 2x sin y + y2 .
∂x x

Example 10.12 Compute the partial derivative with respect to x of f (x, y) =

x 3 y3 .
Solution. We treat y as a constant in our calculation:

∂f
= 3x2 y3 .
∂x
.
Just as the partial derivative of f (x, y) with respect to x was defined, we
can also similarly define the partial derivative of f (x, y) with respect to y at
a point (x0 , y0 ). Let Δy denote a change in y from y0 to y0 + Δy, and keep x
fixed at x0 - see Figure 10.15.
y Figure 10.15: As Δy approaches zero,

the limit of the average rate of change in
the y-direction,
y0 + Δy (x0 , y0 + Δy) f (x0 , y0 + Δy) − f (x0 , y0 )
,
Δy Δy
y0 (x0 , y0 ) is the partial derivative ∂ f /∂ y at (x0 , y0 ).
x
x0
As (x, y) moves parallel to the y-axis from (x0 , y0 ) to (x0 , y0 + Δy), the
average rate of change of f (x, y) is given by:
Δz f (x0 , y0 + Δy) − f (x0 , y0 )
= .
Δy Δz
The limit of this as Δy approaches zero is called the partial derivative

of f (x, y) with respect to y at (x0 , y0 ) and is denoted by:

Δ f f (x0 , y0 + Δy) − f (x0 , y0 )
= lim .
Δy (x0 ,y0 ) Δy→0 Δz
Some other notations used to describe this limit include:

∂ z
and fy (x0 , y0 ).
∂ y (x0 ,y0 )
As (x0 , y0 ) varies throughout the domain of f (x, y) we get a new func-

tion of x and y, denoted by ∂ f /∂ y which is called the partial derivative
of f (x, y) with respect to y.
Example 10.13 Find the partial derivative of f (x, y) = x2 + x3 y2 with respect

to y at the point (x0 , y0 ).
Solution. To compute (∂ f /∂ y)(x0 ,y0 ) we need to take the derivative with

respect to y of the function y:
f (x0 , y) = x02 + x02 y2 ,
which we get from f (x, y) by keeping x fixed at x0 . Thus, because x0 is a

constant, we have
d 2
(x + x03 y2 ) = 2x03 y,
dy 0
and when we replace y with y0 , we obtain

∂ f
= 2x03 y0 .
∂ y (x0 ,y0 )
Example 10.14 Determine the partial derivative with respect to y of f (x, y) =

2
x cos y + y3 ex .
Solution. In order to compute ∂ f /∂ y we keep x fixed and differentiate with

respect to y. This means we treat x as a constant and compute the derivative
with respect to y:
∂f 2
= −x sin y + 3y2 ex .
∂y

Higher order partial derivatives

Partial derivatives of higher order are very important for their applications.
Given a function f (x, y), its partial derivative with respect to x, ∂ f /∂ x, is
another function of x and y - so we may take its partial derivative with respect
to x. This second partial derivative of f (x, y) with respect to x is denoted:

∂ ∂f ∂2 f
= 2,
∂x ∂x ∂x
where on the right hand side we say “second partial of f with respect to x” or
“partial d second f by dx second”. Note that it is also a function of x and y.
Example 10.15 Find the second partial derivative of f with respect to x of
f (x, y) = x3 y2 + xy.
Solution. We keep y fixed and find the first partial derivative with respect
to x:
∂f
= 3x2 y2 + y,
∂x
and then again, keep y fixed and find the second derivative with respect to x:
∂2 f
= 6xy2 .
∂ x2

Perhaps unsurprisingly, we can similarly have the second partial deriva-
tive of f (x, y) with respect to y:

∂ ∂f ∂2 f
= 2.
∂y ∂y ∂y
This is done by differentiating with respect to y twice, keeping x fixed each
time.
Example 10.16 Find the second partial derivative of f (x, y) = x2 ln y2 with

respect to y.
Solution. We differentiate with respect to y twice:
∂f x2 ∂2 f x2
= ; = − .
∂y y ∂ y2 y2

There are also mixed second partial derivatives. This is because
given f (x, y), then ∂ f /∂ x is a function of both x and y, so we may take
it’s partial derivative with respect to y:

∂ ∂f ∂2 f
= .
∂y ∂x ∂ y∂ x
Notice that we still have another partial derivative of f (x, y), that is ∂ f /∂ y
which is also a function of x and y so we can also take its derivative with
respect to x:
∂ ∂f ∂2 f
= .
∂x ∂y ∂ x∂ y
In almost all applications these mixed second partial derivatives are equal,
that is
∂2 f ∂2 f
= .
∂ y∂ x ∂ x∂ y
Thus, the order in which we differentiate is not important. Although we do
not prove it, a theorem exists stating that the mixed second partial derivatives
are equal provided that one mixed derivative is a continuous function of x
and y.
Example 10.17 Compute both mixed partial derivatives of the function
f (x, y) = e3x cos y.
Verify that they are equal.
Solution. Differentiating we have
∂f ∂f
= 3e3x cos y and = −e3x sin y.
∂x ∂y
The mixed second partial derivatives are thus:
∂2 f ∂
= (3e3x cos y) = −3e3x sin y,
∂ y∂ x ∂y
and
∂2 f ∂
= (−e3x sin y) = −3e3x sin y.
∂ x∂ y ∂x
We can see from this that these mixed partial derivatives are equal.
Example 10.18 Consider a closed tube filled with a liquid containing tiny t = t1
particles in suspension. These move randomly so that the molecules of the

substance spread in all directions throughout the liquid – this phenomenon is
Brownian motion (named after the botanist Robert Brown who discovered it
Point of injection
in 1827). x
Suppose further that the liquid itself is not in motion so then Brownian t = t2
motion causes a random process of migration called diffusion which has the
following mathematical description:
Place an x-axis parallel to the axis of the tube, as shown in Fig- Figure 10.16: The particles of a substance
injected into a tube containing a liquid
ure 10.16. For simplicity’s sake, assume that the molecular con- solvent are in random motion, causing the
centration C varies only in the x-direction. The concentration also molecules of the substance to spread in
all directions throughout the liquid. The
depends on the time t. Thus, the concentration is a function of two
distribution of the particles are shown at
independent variables, x and t. Under certain circumstances, the two instants of time after injection at t = 0
concentration C is given as a function of x and t by when 0 < t1 < t2 .
C = t −1/2 e−x
2 /Kt
,
where K is a positive constant (see Figure 10.17).

C C Figure 10.17: Graphs of the function

f (x,t ) = t −1/2 e−x /Kt at two instances
2
of time t1 and t2 where 0 < t1 < t2 . Each

graph is a bell curve, i.e. each is the graph
of a normal distribution (fundamental in
statistics).
t = t1 t = t2
x x
Show that the partial derivatives of this function x and t satisfy
∂C K ∂ 2C
= ,
∂t 4 ∂ x2
which is called a diffusion equation in one dimension.
Solution. We need to compute one partial derivative with respect to t, and

a second partial derivative with respect to x (and perform some arithmetic).
We compute the partial derivative of C with respect to t using the product and
chain rules:
2
∂C 1 x
= − t −3/2 ex /Kt + t −1/2 e−x /Kt
2 2
∂t 2 Kt 2

1 x2
− t −3/2 + t −5/2 e−x /Kt .
2
=
2 K
We compute the partial derivative with respect to x using the chain rule:

∂C −1/2 −x2 /Kt 2x
= t e −
∂x Kt
2
= − t −3/2 xe−x /Kt .
2
Thus, to compute the second partial derivative of C with respect to x, we use

the product and chain rules:

∂ 2C 2 −3/2 −x2 /Kt 2 −3/2 −x2 /Kt −2x
= − t e − t xe
∂ x2 K K Kt
2

2 4x
− t −3/2 + 2 t −5/2 e−x /Kt
2
=
K K

4 1 −3/2 x2 −5/2 −x2 /Kt
= − t + t e .
K 2 K
We identify that this has ∂C/∂t as a factor and obtain:
∂ 2C 4 ∂C
= ,
∂ x2 K ∂t
thus the function C = t −1/2 e−x

2 /Kt
satisfies the diffusion equation.
Diffusion equations may be derived using only physical principles and
experimental facts. The diffusion equation in Example 10.18 governs the
concentration C = C (x,t ) of a substance at point x of a tube at a time t; any
function C = f (x,t ) that expresses the concentration in terms of x and t must
satisfy the diffusion equations, and we have only considered one of many
possible functions.
The diffusion equation is an example of a partial differential equa-
tion. The particular form given in Example 10.18 describes a large class of
phenomena. One such phenomenon is heat conduction in a straight bar of
homogenous material. Indeed, if the x-axis is chosen parallel to the axis of
the bar, then the temperature u at point x and time t is a function of x and t
and satisfies a heat equation,
∂u ∂ 2u
= α2 2 ,
∂t ∂x
where the positive quantity α 2 is a parameter depending only on the material
of the bar. Note the formal similarity between the heat equation and the
diffusion equation. A derivation of the heat equation is based on certain
physical law (for example, Newton’s Law of Cooling).
Partial differential equations similar to diffusion and heat equations have
been applied in the field of ecology to study the migration of a population in a
new environment: by applying the ideas that are used to obtain the diffusion
equation, the dispersion of a population can be studied. The spread of an
infectious disease during an epidemic has also been described by a diffusion
equation of the kind that describes Brownian motion. Without doubt, the
descriptions of the dispersion of a population and spread of a disease are
idealized, because each description is based on the assumption that motion
is random, but such an assumption is necessary as a first step in the study of
these complex phenomena.
10.3 Maximum and minimum values
In Chapter 7 in Math 154, we learned how to find maximum and minimum

values of a function of one variable. In this section we develop similar
techniques for functions of two variables.
Definition 10.2 (Local extrema.) Let f (x, y) be a function defined on a

region R containing the point (x0 , y0 ). Then, f (x, y) has a local maximum at
(x0 , y0 ) if f (x, y) ≤ f (x0 , y0 ) for all points (x, y) that are sufficiently close to
(x0 , y0 ). Also, f (x, y) has a local minimum at (x0 , y0 ) if f (x, y) ≥ f (x0 , y0 )
for all points (x, y) that are sufficiently close to (x0 , y0 ).
Loosely speaking, f (x, y) has a local maximum at (x0 , y0 ) if the point

(x0 , y0 , f (x0 , y0 )) is the highest point on the graph of f (x, y) when compared
with all nearby points. Similarly, f (x, y) has a local minimum at (x0 , y0 ) if
the point (x0 , y0 , f (x0 , y0 )) is temporarily the lowest point on the graph of the
function f (x, y).
Let us examine the behavior of a function f (x, y) near a point (x0 , y0 ) at
which the function has a local extremum. We keep the coordinate y fixed at
y0 and allow x to vary about x0 . Then we obtain that function f (x, y0 ) is a
z
function of one variable x only and its graph is the curve z = f (x, y0 ). If the
Local
y = y0
derivative ∂∂ xf exists, it must be zero: ∂∂ xf (x ,y ) = 0. In the same way, we keep maximum
0 0
the coordinate x fixed at x0 and allow y to vary about y0 . Now the function x = x0
f (x0 , y) is a function of one variable y and its graph is the curve z = f (x0 , y).

If the derivative ∂∂ yf exists, it must be zero: ∂∂ yf (x ,y ) = 0.
0 0
This situation is depicted in Figure 10.18 in the case of a local maximum. y
We say that if a differentiable function has a local extremum at (x0 , y0 ),
then both of the following conditions must be satisfied: (x0 , y0 )

∂ f ∂ f x z = f (x, y0 ) z = f (x0 , y)
= 0 and = 0. (10.3)
∂ x (x0 ,y0 ) ∂ y (x0 ,y0 ) Figure 10.18: Local maximum at (x0 , y0 )
when ∂∂ xf = 0 and ∂∂ yf = 0.
Also, local extrema may happen in such locations on the graph of f (x, y),
z
where the function is not differentiable (Figure 10.19).
(x0 , y0 , f (x0 , y0 ))
Definition 10.3 A critical point of f (x, y) is a point (x0 , y0 ) in the domain of
the function f such that both

∂ f ∂ f
= 0 and = 0,
∂ x (x0 ,y0 ) ∂ y (x0 ,y0 ) y
or at least one of the partial derivatives does not exist.

(x0 , y0 )
We can deduce that local extrema occur at critical points, just as for a x
function of one variable. However, recall that for a function of one variable a Figure 10.19: Local maximum at (x0 , y0 )
local extremum need not occur at every critical point. For example, f (x) = x3 when ∂∂ xf and ∂∂ yf do not exist.
has an inflection point at the critical point x = 0. Similar situations may
occur in two-dimensional analysis. Consider the function f (x, y) = y2 − x2
(Figure 10.20). This function has first partial derivatives ∂∂ xf (x, y) = −2x and
∂f
∂ y (x, y) = 2y from which it follows that

∂ f ∂ f
= 0 and = 0.
∂ x (0,0) ∂ y (0,0)
z
Thus, (0, 0) is a critical point. However, it is clear from the graph that it is a
temporarily highest point on the surface if we move in the x-direction to the
left and to the right. At the same time, it is a temporarily lowest point on the
surface if we move in the y-direction. Thus, the function f (x, y) = y2 − x2 does (0, 0)
not have a local extremum at (0, 0). Because of the resemblance of the shape y
to a saddle for a horse, the point (0, 0) is called a saddle point.

To determine the nature of a critical point of a function f (x, y) of two z = y 2 − x2
variables, we use the second derivatives of f (x, y). The resulting test, which x
Figure 10.20: The function f (x, y) = y2 − x2

has neither a local maximum nor a local
minimum at the critical point (0, 0).
helps us classify these points, is called the second derivative test. Note that
we can apply this test only at such critical points where all first and second
derivatives exist.
Summary: second derivative test.
1. Find the critical points of f (x, y) by solving the system of simultaneous

equations
∂ f ∂ f
= 0 and = 0.
∂x (x0 ,y0 ) ∂y (x0 ,y0 )
2
∂2 f
· ∂∂ y2f − ∂2 f
2
2. Let D(x, y) = ∂ x2 ∂ x∂ y . Then,
(a) If D(x0 , y0 ) > 0 and fxx (x0 , y0 ) > 0, then f (x, y) has a local minimum
at (x0 , y0 ).
(b) If D(x0 , y0 ) > 0 and fxx (x0 , y0 ) < 0, then f (x, y) has a local maximum
at (x0 , y0 ).
(c) If D(x0 , y0 ) < 0, then (x0 , y0 ) is a saddle point.
(d) If D(x0 , y0 ) = 0, then the test is inconclusive and some other techniques
must be used.
Example 10.19 Find the local extrema of the function f (x, y) = 2x2 + y2 + 1.
Solution. Compute the first partial derivatives
∂f ∂f
= 4x, = 2y.
∂x ∂y
Clearly, (0, 0) is the critical point and it is the only point that satisfies both
equations.
Now find the second partial derivatives
∂2 f ∂2 f ∂2 f
= 4, = 2, and = 0.
∂ x2 ∂ y2 ∂ x∂ y

∂2 f
Therefore, D(x, y) = 4 · 2 − 0 = 8 and D(0, 0) = 8 > 0. Also, ∂ x2 (0,0)
> 0.
Based on the second derivative test, we conclude that f (x, y) = 2x + y2 + 1
2
has a local minimum at (0, 0).

Example 10.20 Find the local extrema of the function
f (x, y) = −x2 + xy − y2 − 5x.
Solution. The first partial derivatives are
∂f ∂f
= −2x + y − 5 and = x − 2y,
∂x ∂y
and so the local extreme values occur at those points that satisfy the simulta-
neous equations
−2x + y − 5 = 0,
and
x − 2y = 0.
From the second equation we have x = 2y, which we substitute into the
first equation to obtain −2(2y) + y − 5 = 0 and, then solving for y, we have
y = −5/3. As x = 2y, it follows that x = −10/3. Thus the only point where
a local extreme value may occur is at (−10/3, −5/3). Now we apply the
second derivative test. Computing the second derivatives, we have
∂2 f ∂2 f ∂2 f
= −2, = −2, and =1 for all values of x and y.
∂ x2 ∂ y2 ∂ x∂ y
Thus
2 2
∂2 f ∂2 f ∂ f
D(x, y) = 2 2 − = (−2)(−2) − 1 = 3 > 0,
∂x ∂y ∂ x∂ y
and
∂2 f
= −2 < 0.
∂ x2
Hence, by the second derivative test we conclude that f (−10/3, −5/3) =
75/9 is a local maximum value. There are no other local extreme values
because (−10/3, −5/3) is the only point where both first partial derivatives
are zero.

0
−10
Example 10.21 Locate all local extrema and saddle points of

−20
z
f (x, y) = 4xy − x − y .
4 4
−30
−40
Solution. Since
∂f −2
= 4y − 4x3 , −1
−2
∂x y 0 0
−1
(10.4) 1 x
∂f
1
2 2
= 4x − 4y3 ,
∂y Figure 10.21: Graph of the function
f (x, y) = 4xy − x4 − y4 .
the critical points of f have coordinates satisfying the equations
4y − 4x3 = 0, or y = x3 , (10.5)
4x − 4y = 0,
3
or x=y . 3
(10.6)
Substituting Eqn. (10.5) into Eqn. (10.6) yields x = (x3 )3 , or equivalently

x9 − x = 0 or x(x8 − 1) = 0, which has solutions x = 0, x = 1, x = −1.
Substituting these values in Eqn. (10.5), we obtain the corresponding y-values
y = 0, y = 1, y = −1. Thus, the critical points of f are (0, 0),(1, 1), and
(−1, −1).
From Eqn. (10.4)

∂2 f ∂2 f ∂2 f
= −12x2 , = −12y2 , = 4,
∂ x2 ∂ y2 ∂ x∂ y
which yields the following table:

CRITICAL POINT ∂2 f ∂2 f ∂2 f Table 10.3: Critical points and partial

∂ x2 (x ,y )

∂ y2 (x ,y ) ∂ x∂ y (x ,y ) D(x0 , y0 ) derivatives of the function f (x, y) =
(x0 , y0 ) 0 0 0 0 0 0
4xy − x4 − y4 .
(0, 0) 0 0 4 −16
(1, 1) −12 −12 4 128
(−1, −1) −12 −12 4 128 3
2
Since at the points (1, 1) and (−1, −1), D(x0 , y0 ) > 0 and ∂∂ x2f (x ,y ) < 0,
2
0 0
we conclude that local maxima occur at these critical points. At (0, 0) there
1
is a saddle point since D(0, 0) < 0. The graph of the function is given in 0
Figure 10.21, the contour plot is presented in Figure 10.22. -1
-2
Example 10.22 The Shannon index (sometimes called the Shannon-Wiener -3

-3 -2 -1 0 1 2 3
index or Shannon-Weaver index) is a measure of diversity in an ecosystem. Figure 10.22: Contour plot of the function
For the case of three species, it is defined as f (x, y) = 4xy − x4 − y4 .
H = −p1 ln p1 − p2 ln p2 − p3 ln p3 ,
where pi is the proportion of the ecosystem made up of species i.
(a) Express H as a function of two variables using the fact that
p1 + p2 + p3 = 1.
(b) What is the domain of H?
(c) Find the maximum value of H. For what values of p1 , p2 , p3 does it occur?
Solution.
(a) Given that p1 + p2 + p3 = 1, we have p3 = 1 − p1 − p2 . Then
H ( p1 , p2 ) = −p1 ln p1 − p2 ln p2 − (1 − p1 − p2 ) ln(1 − p1 − p2 ).
(b) To find the domain of the function H ( p1 , p2 ), we recall that since pi p2

(i = 1, 2, 3) is a proportion of species i, 0 < pi < 1, that is, 0 < p1 < 1,
0 < p2 < 1, and 0 < p3 < 1, or 0 < 1 − p1 − p2 < 1. Solving the last
inequality leads to 1
0 < p1 + p2 < 1. p1 + p 2 = 1
Note that we excluded values pi = 0 and pi = 1 (i = 1, 2, 3).
We therefore conclude that the set of points which satisfies all three
inequalities (0 < p1 < 1, 0 < p2 < 1, and 0 < p1 + p2 < 1), is the interior
0 1 p1
of the triangle shown in Figure 10.23, excluding the boundaries.
Figure 10.23: The domain of H ( p1 , p2 ) in
Example 10.22
(c) We must find the partial derivatives

∂H
= − ln p1 − 1 + ln(1 − p1 − p2 ) + 1
∂ p1
1 − p1 − p2
= ln ,
p1
∂H
= − ln p2 − 1 + ln(1 − p1 − p2 ) + 1
∂ p2
1 − p1 − p2
= ln ,
p2
∂H ∂H
To find critical points, we set ∂ p1 = 0 and ∂ p2 = 0. Thus
1 − p1 − p2 1 − p 1 − p2
ln =0 and ln = 0, or
p1 p2
1 − p1 − p 2 1 − p 1 − p2
=1 and ln = 1.
p1 p2
Hence,
1 − p1 − p2 = p1 , 1 − p1 − p2 = p2 ,
1 − 2p1 = p2 , 1 − 2p2 = p1 .
Substituting p1 from the last equation into the first one, we obtain
1 − 2(1 − 2p2 ) = p2 ,
−1 + 4p2 = p2 ,
p2 = 1/3.
Finally, p1 = 1 − 2p2 = 1/3 (p3 = 1 − p1 − p2 = 1/3). The only

critical point of the function H ( p1 , p2 ) is (1/3, 1/3). To classify this
critical point, we find the second partial derivatives and apply the second
derivative test. Thus,
∂ 2H p1 −p1 − (1 − p1 − p2 ) p2 − 1
= · = ,
∂ p1
2 1 − p1 − p2 2
p1 p1 (1 − p1 − p2 )
∂ 2H p2 −p2 − (1 − p1 − p2 ) p1 − 1
= · = ,
∂ p22 1 − p1 − p2 p22 p2 (1 − p1 − p2 )

∂ 2H p1 1 1
= · − =− .
∂ p1 ∂ p2 1 − p1 − p2 p1 1 − p1 − p2
Then

∂ 2 H ∂ 2 H ∂ 2 H
= −6, = −6, = −3,
∂ p21 ( 1 , 1 ) ∂ p22 ( 1 , 1 ) ∂ p1 ∂ p2 ( 1 , 1 )
3 3 3 3 3 3
and
1 1
D( , ) = (−6)(−6) − (−3)2 = 27 > 0.
3 3
This leads to the conclusion that H ( p1 , p2 ) achieves its local maximum
at (1/3, 1/3): H (1/3, 1/3) = ln 3. Since it is the only extremum and the
domain of H ( p1 , p2 ) is unbounded, we can call it the global maximum as
it was done in Math 154.
10.4 The chain rule
We studied the chain rule in Chapter 9 in Math 154. Recall that if y = g(u) and
u = f (x) are both differentiable functions and y = g( f (x)) is the composite
function, then the chain rule of differentiation states that
dy dy du
= · .
dx du dx
Assume that z = f (x, y) is a function of x and y, and suppose that x and y are
in turn functions of the single variable t:
x = x(t ), y = y(t ).
Then the composite function z = f (x(t ), y(t )) is a function of the single

variable t.
The chain rule of differentiation z
dz ∂ z dx ∂ z dy
= + . (10.7) ∂z ∂z
dt ∂ x dt ∂ y dt ∂x ∂y
Formula Eqn. (10.7) is represented schematically in Figure 10.24. Note y
x
that we do not prove it here as the proof is beyond the scope of this course.
dx dy
Example 10.23 Suppose that dt dt
t t
z = x2 y, x = t 2, y = t 3.
Figure 10.24: The chain rule in the case of
z = f (x, y) and x = x(t ), y = y(t ).
Use the chain rule to find dz/dt, and check the result by expressing z as a
function of t and differentiating directly.
Solution. By the chain rule

= + = (2xy)(2t ) + (x2 )(3t 2 )
dt ∂ x dt ∂ y dt
= (2t 5 )(2t ) + (t 4 )(3t 2 ) = 7t 6 .
Alternatively, we can express z directly as a function of t,
z = x2 y = (t 2 )2 (t 3 ) = t 7 ,
and then differentiate to obtain dz/dt = 7t 6 . However, this procedure may not
always be convenient.

Example 10.24 Suppose that

√
z= xy + y, x = cos θ , y = sin θ .
Use the chain rule to find dz/dθ when θ = π/2.

Solution. From the chain rule with θ in place of t,

= + ,
dθ ∂ x dθ ∂ y dθ
we obtain
dz 1 1
= (xy + y)−1/2 (y)(− sin θ ) + (xy + y)−1/2 (x + 1)(cos θ ).
dθ 2 2
When θ = π/2, we have
π π
x = cos = 0, y = sin = 1.
2 2
Substituting x = 0, y = 1, θ = π/2 in the formula for dz/dθ yields

dz 1 1 1
= (1)(1)(−1) + (1)(1)(0) = − .
dθ θ =π/2 2 2 2
Example 10.25 (Related rates) At what rate is the area of the rectangle
changing if its length is 8 m and increasing at a rate of 3 m/s, and its width is
6 m and is decreasing at a rate of 2 m/s?
Solution. Let x and y denote the length and the width of the rectangle, t –
time in seconds, and S – area of the rectangle. Then
dx dy
S = xy, = 3, = −2.
dt dt
Applying the chain rule:
dS ∂ S dx ∂ S dy
= +
dt ∂ x dt ∂ y dt
= y · 3 + x · (−2) = 6 · 3 − 8 · 2 = 18 − 16 = 2(m2 /s).

Two-variable chain rule
If x = x(u, v) and y = y(u, v) have first-order partial derivatives at the point
(u, v), and if z = f (x, y) is differentiable at the point (x(u, v), y(u, v)), then z =
f (x(u, v), y(u, v)) has first-order partial derivatives at (u, v) given by
∂z ∂z ∂x ∂z ∂y ∂z ∂z ∂x ∂z ∂y
= + and = + . (10.8)
∂u ∂x ∂u ∂y ∂u ∂v ∂x ∂v ∂y ∂v
The two-variable chain rule, Eqn. (10.8), is a natural extension of
Eqn. (10.7).
Example 10.26 Given that
z = exy , x = 2u + v, y = u/v,
∂z ∂z
find ∂u and ∂v using the chain rule.
Solution.

∂z ∂z ∂x ∂z ∂y 1 x
= + = (yexy )(2) + (xexy ) = 2y + exy
∂u ∂x ∂u ∂y ∂u v v

2u 2u + v (2u+v)(u/v) 4u
= + e = + 1 e(2u+v)(u/v) ,
v v v
∂z ∂z ∂x ∂z ∂y u
= + = (yexy )(1) + (xexy ) − 2
∂v ∂x ∂v ∂y ∂v v
u u u
= y−x 2 e = xy
− (2u + v) 2 e(2u+v)(u/v)
v v v
2u2 (2u+v)(u/v)
=− 2 e .
v

Exercises
10.1. Computing values. Compute the values of f (x, y) at the following

points (0, 0); (1, 1); (2, 1), (1, 2); (−1, 1); (2, −1); (−3, −3).
(a) f (x, y) = 2x + y.
(b) f (x, y) = x + y2 .
(c) f (x, y) = 3xy + 2.
(d) f (x, y) = x2 y + y.
10.2. Sketching contour lines. Sketch the contour line f (x, y) = c for each
of the given values of the constant c.
(a) f (x, y) = x + 2y; c = 0, 1, 2.
(b) f (x, y) = 2x + y; c = 0, 1, 2.
(c) f (x, y) = x2 + y2 + 1; c = 2, 5, 10.
(d) f (x, y) = 1 + y2 ; c = 1, 2, 3.
(e) f (x, y) = 3x2 + 3y2 ; c = 1, 3, 5.
(f) f (x, y) = −x2 − y2 ; c = −1, −2, −3.
(g) f (x, y) = xy; c = 1, 2, 3.
(h) f (x, y) = 2x2 + y2 ; c = 4, 6, 8.
(i) f (x, y) = 3x2 + 4y2 c = 3, 6, 9.
(j) f (x, y) = x2 − y2 c = 0, 1, 2.
(k) f (x, y) = x1/2 y c = 1, 2, 3.
(l) f (x, y) = e−x − 2y c = 2, 4, 6.
10.3. Graph intersections. Sketch the graph of the function f (x, y) = 2x2 +
2y2 . Sketch the intersection of the graph with the plane z = 4. Describe
the relation between this intersection and the contour line f (x, y) = 4.
10.4. Graph intersections. Sketch the graph of the function f (x, y) = 2x + y,

showing the intersection of the graph with the three planes x = 0, y = 0
and x = 3.
10.5. Graph intersections. Sketch the graph of the function f (x, y) =
3 − x2 − y2 , showing the intersections of the graph with the planes x =
0, y = 0 and z = 0.
10.6. Graph intersections. Sketch the graph of the function f (x, y) = 5 − x2 ,
showing the intersections of the graph with the planes x = 0, y = 0, z =
0 and z = 4.
10.7. Graph intersection. Sketch the graph of the function f (x, y) = (2x −
4)2 . Show the intersection of the graph with each of the planes x = 1
and y = 1.
10.8. Contour line through point. For each designated function f (x, y),
sketch the contour line that passes through the point (a, b).
(a) f (x, y) = 4x2 − 2y + 3; (a, b) = (1, 3).

(b) f (x, y) = ye−x ; (a, b) = (0, 1).
(c) f (x, y) = ( x − 1 ) 2 + y2 ; (a, b) = (3, 0).
sin x π
(d) f (x, y) = ; (a, b) = ( , 3).
y 2
10.9. Partial derivatives. For each function f (x, y), find the partial deriva-
tives ∂ f /∂ x and ∂ f /∂ y.
(a) f (x, y) = 3x2 + y2 .
(b) f (x, y) = 2x + 3y.
(c) f (x, y) = exy .
(d) f (x, y) = xey .
(e) f (x, y) = sin(xy).
(f) f (x, y) = ln(x2 + y2 ).
(g) f (x, y) = xy2 .
x2
(h) f (x, y) = .
y3
(i) f (x, y) = ey cos x.
(j) f (x, y) = y sin x.
(k) f (x, y) = xy.
x
(l) f (x, y) = .
x+y
xy
(m) f (x, y) = 2 .
x + y2
(n) f (x, y) = ex+y .
x2 − y2
(o) f (x, y) = .
x2 − y2
(p) f (x, y) = y−1/2 e−x

2 /y
.
(q) f (x, y) = (x2 + y2 )2 .
1 1
(r) f (x, y) = − .
x y
(s) f (x, y) = xy(3 − x − y).

x
(t) f (x, y) = ln .
y
10.10. Partial derivatives at a point. For each function f (x, y), find the par-
tial derivative (∂ f /∂ x)(x0 ,y0 ) and (∂ f /∂ y)(x0 ,y0 ) , at each designated
point (x0 , y0 ).
(a) f (x, y) = x2 + 3y2 , (x0 , y0 ) = (2, 1).
(b) f (x, y) = 5x + 7y, (x0 , y0 ) = (1, 1).
(c) f (x, y) = e2xy , (x0 , y0 ) = (0, 1).
(d) f (x, y) = ln(x2 + y2 ), (x0 , y0 ) = (1, 0).
(e) f (x, y) = x2 y, (x0 , y0 ) = (1, 2).
π
(f) f (x, y) = sin(2xy), (x0 , y0 ) = 1, .
2

x
(g) f (x, y) = ln (x0 , y0 ) = (e, 1).
y
2x
(h) f (x, y) = , (x0 , y0 ) = (−1, −2).
x 2 + y2
(i) f (x, y) = 3ex+y , (x0 , y0 ) = (1, −1).
π
(j) f (x, y) = cos(x + y), (x0 , y0 ) = − , 2π .
2
10.11. Second partial derivatives. For each of the following functions f (x, y),
find the second partial derivatives, ∂ 2 f /∂ x2 , ∂ 2 f /∂ y2 , ∂ 2 f /∂ y∂ x,
and ∂ 2 f /∂ x∂ y. Verify that the two mixed second partial deriva-
tives, ∂ 2 f /∂ y∂ x and ∂ 2 f /∂ x∂ y, are equal.
(a) f (x, y) = xy2 + x2 .

(b) f (x, y) = ye3x .
y
(c) f (x, y) = 2 .
x + y2
(d) f (x, y) = e3x+3y .
(e) f (x, y) = sin(x + y).
(f) f (x, y) = x cos y.
(g) f (x, y) = ln(x2 + y2 )1/2 .
y3
(h) f (x, y) = .
x2
(i) f (x, y) = sin(2x) cos(3y).
1
(j) f (x, y) = (e2x + e−2x ) sin(2y).
2
10.12. Wave equation. Let a and b denote constants. Show that the function
u(x, y) = sin(ax) cos(bt )
of x and t satisfies a one-dimensional wave equation
∂ 2 u a 2 ∂ 2 u
= .
∂ x2 b ∂t 2
10.13. Laplace equation. Let a denote a constant. Show that the function
1
v(x, y) = (eax − e−ax ) sin(ay)
2
satisfies the Laplace equation in two dimensions:
∂ 2v ∂ 2v
+ = 0.
∂ x2 ∂ y2
10.14. Extreme values. Find all extreme values of the function f (x, y).
(a) f (x, y) = x2 + 2y2 − 4x + 6y − 2. (g) f (x, y) = x3 + y2 − 6xy + 6x + 3y − 7.

(b) f (x, y) = x2 + y4 + 3x − y2 − 4. (h) f (x, y) = x3 + 3xy2 − 3x2 − 3y2 + 1.
1 1
(c) f (x, y) = x3 + y3 − 9xy + 27. (i) f (x, y) = xy + + .
x y
(d) f (x, y) = x3 + 3xy2 − 15x + y3 − 15y. (j) f (x, y) = x2 + 2y − y2 − 4y + 3.
(e) f (x, y) = x2 + y3 − 6y. (k) f (x, y) = (x − 3)2 + (y + 5)2 + 2.
(f) f (x, y) = x2 − 3x + xy + 2y2 − y + 5. (l) f (x, y) = x3 + y3 .
10.15. Chain rule. Use the chain rule to find dz

dt of the function z = f (x, y).
(a) z = xy; x = t, y = t 2 .
(b) z = x2 + 3y2 ; x = t 1/2 , y = t + 1.
(c) z = ln(x2 + y2 ); x = 1 + t 2, y = t 3.
(d) z = ex+y ; x = t 2 , y = 4t.
(e) z = 2x + 3y; x = cos(5t ), y = sin(5t ).
(f) z = ex cos y; x = 2 + 3t, y = t 2 .
ex + y
(g) z = ; x = 1 + 2t, y = −1 + t 2 .
1 + x2
1
(h) z = exy ; x = t 2 , y = .
t
(i) z = x + y ; x = cos(3t ), y = sin(3t ).
2 2
(j) z = 4x2 + y2 ; x = cos t, y = 2 sin t.

(k) z = x2 − y2 ; x = et + e−t , y = et − e−t .
(l) z = ex sin y ; x = 5t, y = t 2 .
10.16. Chain rule. Use the chain rule to find ∂ z/∂ u and ∂ z/∂ v.
(a) z = 8x2 y − 2x + 3y; x = uv, y = u − v.

(b) z = x2 − y tan x; x = u/v, y = u2 v2 .
(c) z = x/y; x = 2 cos u, y = 3 sin v.
(d) z = 3x − 2y; x = u + v ln u, y = u2 − v ln v.
x2 y √
(e) z = e ; x = uv, y = 1/v.
(f) z = cos x sin y; x = u − v, y = u2 + v2 .
10.17. Chain rule. Let T = x2 y − xy3 + 2; x = r cos θ , y = r sin θ . Find

∂ T /∂ r and ∂ T /∂ θ .

CNQ#10 PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

CNQ#10 PDF

Încărcat de

Drepturi de autor:

Formate disponibile

10

So far we have only considered functions of one independent variable -

10.1 Functions of two or more independent variables

• let w denote weight measured in kilograms,

s = (0.007184)w0.425 h0.725 . (10.1)

f (w, h) = (0.007184)w0.425 h0.725 .

The relationship s = f (w, h) is accurate for values of weight w from 20 kg

s = f (35, 130) = (0.007184)(350.425 )(1300.725 ) = 1.1097 m2

deﬁnes z as a function f of the independent variables x and y if, for each

f (x, y) = x + y. Figure 10.1: Domain D = {(w, h) : 20 ≤ w ≤

Solution. The domain D is pictured in Figure 10.2. To ﬁnd the range

f (1, 0) = 1, f (0, 1) = 1, f (0.5, 0.5) = 1, f (0.2, 0.8) = 1,

140 chart is used by physiologists in metabolism

Example 10.5 We return to Example 10.4. Let D = {(x, y) : 0 ≤ x ≤ 1, 0 ≤

Plot the contour line such that f (x, y) = 1.

(x, y, f (x, y))

is a possible value of an independent variable in the domain of f (x).

xyz-space. For every point (x0 , y0 ) in the domain D of a function z =

Figure 10.7: The graph of the func-

The graph of the function f (x, y) = x2 + y2 in Example 10.6 is shown in

Level curves have wide applications in various ﬁelds. One common

example of level curves occurs in topographic maps of mountainous regions,

Figure 10.9: A topographical map with level

is a plane if a, b, and c are not all zero.

Example 10.7 Graph 2x + y + z = 8.

z Figure 10.10: The triangular portion of the

z Figure 10.12: The graph of f (x, y) = 3 − y2

The next example concerns a function with elliptic contour lines.

10.2 Partial derivatives and applications

In order to introduce partial derivatives, let us ﬁrst suppose that z = f (x, y) is a

Δz f (x0 + Δx, y0 ) − f (x0 , y0 )

Deﬁnition 10.1 The instantaneous rate of change of f (x, y) in the x-direction

Δz f (x0 + Δx, y0 ) − f (x0 , y0 )

y Figure 10.14: As a point (x, y) moves

Note: the symbol ∂ in the partial derivative ∂ f /∂ x is a rounded “d” and we

Example 10.10 Let f (x, y) = x2 + x3 y2 . Find its partial derivative with

Solution. Referring to the deﬁnition of (∂ f /∂ x)(x0 ,y0 ) , we note that we are

Note: we of course can ﬁnd the partial derivative of f (x, y) (described

if (x0 , y0 ) = (2, 3), then

Example 10.11 Find the partial derivative with respect to x of f (x, y) =

with respect to x by treating y as a constant in our calculations. Thus, we get:

Example 10.12 Compute the partial derivative with respect to x of f (x, y) =

Solution. We treat y as a constant in our calculation:

y Figure 10.15: As Δy approaches zero,

The limit of this as Δy approaches zero is called the partial derivative

Some other notations used to describe this limit include:

As (x0 , y0 ) varies throughout the domain of f (x, y) we get a new func-

Example 10.13 Find the partial derivative of f (x, y) = x2 + x3 y2 with respect

Solution. To compute (∂ f /∂ y)(x0 ,y0 ) we need to take the derivative with

f (x0 , y) = x02 + x02 y2 ,

which we get from f (x, y) by keeping x ﬁxed at x0 . Thus, because x0 is a

Example 10.14 Determine the partial derivative with respect to y of f (x, y) =

Solution. In order to compute ∂ f /∂ y we keep x ﬁxed and differentiate with

Higher order partial derivatives

Example 10.15 Find the second partial derivative of f with respect to x of

Example 10.16 Find the second partial derivative of f (x, y) = x2 ln y2 with

Solution. We differentiate with respect to y twice:

Example 10.17 Compute both mixed partial derivatives of the function

f (x, y) = e3x cos y.

Verify that they are equal.

Solution. Differentiating we have

The mixed second partial derivatives are thus:

particles in suspension. These move randomly so that the molecules of the