Copyright © Cengage Learning. All rights reserved.

**14.3** Partial Derivatives

### Partial Derivatives

On a hot day, extreme humidity makes us think the

temperature is higher than it really is, whereas in very dry air we perceive the temperature to be lower than the

thermometer indicates.

*The National Weather Service has devised the heat index *
(also called the temperature-humidity index, or humidex, in
some countries) to describe the combined effects of

temperature and humidity.

*The heat index I is the perceived air temperature when the *
*actual temperature is T and the relative humidity is H.*

*So I is a function of T and H and we can write I = f(T, H).*

3

### Partial Derivatives

*The following table of values of I is an excerpt from a table *
compiled by the National Weather Service.

**Table 1**

*Heat index I as a function of*
temperature and humidity

### Partial Derivatives

If we concentrate on the highlighted column of the table,
*which corresponds to a relative humidity of H = 70%, we *
are considering the heat index as a function of the single
*variable T for a fixed value of H. Let’s write g(T ) = f(T, 70). *

*Then g(T ) describes how the heat index I increases as the *
*actual temperature T increases when the relative humidity *
is 70%.

*The derivative of g when T = 96°F is the rate of change of *
*I with respect to T when T = 96°F:*

5

### Partial Derivatives

*We can approximate g*′(96) using the values in Table 1 by
*taking h = 2 and –2:*

Averaging these values, we can say that the derivative
*g*′(96) is approximately 3.75.

### Partial Derivatives

This means that, when the actual temperature is 96°F and the relative humidity is 70%, the apparent temperature

(heat index) rises by about 3.75°F for every degree that the actual temperature rises!

7

### Partial Derivatives

Now let’s look at the highlighted row in Table 1, which
*corresponds to a fixed temperature of T = 96°F.*

**Table 1**

*Heat index I as a function of*
temperature and humidity

### Partial Derivatives

The numbers in this row are values of the function
*G(H) = f(96, H), which describes how the heat index *

*increases as the relative humidity H increases when the *
*actual temperature is T = 96°F.*

*The derivative of this function when H = 70% is the rate of *
*change of I with respect to H when H = 70%:*

9

### Partial Derivatives

*By taking h = 5 and –5, we approximate G′(70) using the *
tabular values:

By averaging these values we get the estimate

*G′(70) ≈ 0.9. This says that, when the temperature is 96°F*
and the relative humidity is 70%, the heat index rises about
0.9°F for every percent that the relative humidity rises.

### Partial Derivatives

*In general, if f is a function of two variables x and y,*

*suppose we let only x vary while keeping y fixed, say y = b, *
*where b is a constant. *

Then we are really considering a function of a single

*variable x, namely, g(x) = f(x, b). If g has a derivative at a, *
**then we call it the partial derivative of f with respect to x ****at (a, b) and denote it by f**_{x}*(a, b). Thus*

11

### Partial Derivatives

By the definition of a derivative, we have

and so Equation 1 becomes

### Partial Derivatives

**Similarly, the partial derivative of f with respect to y at****(a, b), denoted by f**_{y}*(a, b), is obtained by keeping x fixed*
*(x = a) and finding the ordinary derivative at b of the *

*function G(y) = f(a, y):*

With this notation for partial derivatives, we can write the
*rates of change of the heat index I with respect to the *

*actual temperature T and relative humidity H when *
*T = 96°F and H = 70% as follows:*

*f** _{T}*(96, 70) ≈ 3.75 f

*(96, 70) ≈ 0.9*

_{H}13

### Partial Derivatives

*If we now let the point (a, b) vary in Equations 2 and 3, *
*f*_{x}*and f** _{y}* become functions of two variables.

### Partial Derivatives

There are many alternative notations for partial derivatives.

*For instance, instead of f*_{x}*we can write f*_{1} or D_{1}*f (to indicate *
*differentiation with respect to the first variable) or ∂f/∂x. *

But here *∂f/∂x can’t be interpreted as a ratio of differentials.*

15

### Partial Derivatives

To compute partial derivatives, all we have to do is

remember from Equation 1 that the partial derivative with
*respect to x is just the ordinary derivative of the function g*
*of a single variable that we get by keeping y fixed. *

Thus we have the following rule.

*If f(x, y) = x*^{3} *+ x*^{2}*y*^{3} *– 2y*^{2}*, find f*_{x}*(2, 1) and f** _{y}*(2, 1).

Solution:

*Holding y constant and differentiating with respect to x, *
we get

*f*_{x}*(x, y) = 3x*^{2} *+ 2xy*^{3}

and so *f** _{x}*(2, 1) = 3 2

^{2}+ 2 2 1

^{3 }

*Holding x constant and differentiating with respect to y, *
we get

*f*_{y}*(x, y) = 3x*^{2}*y*^{2} *– 4y*

*f** _{y}*(2, 1) = 3 2

^{2} 1

^{2}– 4 1

### Example 1

= 16

= 8

17

### Interpretations of Partial Derivatives

### Interpretations of Partial Derivatives

To give a geometric interpretation of partial derivatives, we
*recall that the equation z = f(x, y) represents a surface*

*S (the graph of f ). If f(a, b) = c, then the point P(a, b, c) lies *
*on S.*

*By fixing y = b, we are restricting our attention to the curve *
*C*_{1} *in which the vertical plane y = b intersects S. (In other *
*words, C*_{1} *is the trace of S in the plane y = b.)*

19

### Interpretations of Partial Derivatives

*Likewise, the vertical plane x = a intersects S in a curve C*_{2}.
*Both of the curves C*_{1} *and C*_{2} *pass through the point P. *

(See Figure 1.)

*Note that the curve C*_{1} is the
graph of the function

*g(x) = f(x, b), so the slope of its *
*tangent T*_{1} *at P is g′(a) = f*_{x}*(a, b). *

*The curve C*_{2} is the graph of the
*function G(y) = f(a, y), so the *
*slope of its tangent T*_{2} *at P is*
*G′(b) = f*_{y}*(a, b). *

**Figure 1**

*The partial derivatives of f at (a, b) are*
*the slopes of the tangents to C*_{1}*and C*_{2}.

### Interpretations of Partial Derivatives

*Thus the partial derivatives f*_{x}*(a, b) and f*_{y}*(a, b) can be *

interpreted geometrically as the slopes of the tangent lines
*at P(a, b, c) to the traces C*_{1} *and C*_{2} *of S in the planes y = b*
*and x = a. *

As we have seen in the case of the heat index function,
*partial derivatives can also be interpreted as rates of *
*change. *

*If z = f(x, y), then ∂z/∂x represents the rate of change of z *
*with respect to x when y is fixed. Similarly, ∂z/∂y*

*represents the rate of change of z with respect to y when x*
is fixed.

21

### Example 2

*If f(x, y) = 4 – x*^{2} *– 2y*^{2}*, find f*_{x}*(1, 1) and f** _{y}*(1, 1) and interpret
these numbers as slopes.

Solution:

We have

*f*_{x}*(x, y) = –2x f*_{y}*(x, y) = –4y*
*f*_{x}*(1, 1) = –2 f** _{y}*(1, 1) = –4
.

*Example 2 – Solution*

*The graph of f is the paraboloid z = 4 – x*^{2} *– 2y*^{2} and the

*vertical plane y = 1 intersects it in the parabola z = 2 – x*^{2},
*y = 1. (As in the preceding discussion, we label it C*_{1} in

Figure 2.)

The slope of the tangent line to this parabola at the point

*(1, 1, 1) is f** _{x}*(1, 1) = –2.

**Figure 2**

cont’d

23

*Example 2 – Solution*

*Similarly, the curve C*_{2} *in which the plane x = 1 intersects *
*the paraboloid is the parabola z = 3 – 2y*^{2}*, x = 1, and the *
*slope of the tangent line at (1, 1, 1) is f** _{y}*(1, 1) = –4. (See
Figure 3.)

**Figure 3**

cont’d

### Functions of More Than Two

### Variables

25

### Functions of More Than Two Variables

Partial derivatives can also be defined for functions of three
*or more variables. For example, if f is a function of three *

*variables x, y, and z, then its partial derivative with respect *
*to x is defined as*

*and it is found by regarding y and z as constants and *
*differentiating f(x, y, z) with respect to x.*

### Functions of More Than Two Variables

*If w = f(x, y, z), then f** _{x}* =

*∂w/∂x can be interpreted as the*

*rate of change of w with respect to x when y and z are held*fixed. But we can’t interpret it geometrically because the

*graph of f lies in four-dimensional space.*

*In general, if u is a function of n variables, *

*u = f(x*_{1}*, x*_{2}*,…, x** _{n}*), its partial derivative with respect to the

*i th variable x*

*is*

_{i}and we also write

27

### Example 6

*Find f*_{x}*, f*_{y}*, and f*_{z}*if f(x, y, z) = e*^{xy}*ln z.*

Solution:

*Holding y and z constant and differentiating with respect *
*to x, we have*

*f*_{x}*= ye*^{xy}*ln z*
Similarly,

*f*_{y}*= xe*^{xy}*ln z and f** _{z}* =

### Higher Derivatives

29

### Higher Derivatives

*If f is a function of two variables, then its partial derivatives *
*f*_{x}*and f** _{y}* are also functions of two variables, so we can

*consider their partial derivatives (f** _{x}*)

_{x}*, (f*

*)*

_{x}

_{y}*, (f*

*)*

_{y}

_{x}*, and (f*

*)*

_{y}*,*

_{y}

**which are called the second partial derivatives of f.***If z = f(x, y), we use the following notation:*

### Higher Derivatives

*Thus the notation f** _{xy }*(or ∂

^{2}

*f/∂y ∂x) means that we first*

*differentiate with respect to x and then with respect to y,*

*whereas in computing f*

*the order is reversed.*

_{yx}31

### Example 7

Find the second partial derivatives of
*f(x, y) = x*^{3} *+ x*^{2}*y*^{3} *– 2y*^{2}

Solution:

In Example 1 we found that

*f*_{x}*(x, y) = 3x*^{2} *+ 2xy*^{3} *f*_{y}*(x, y) = 3x*^{2}*y*^{2} *– 4y*
Therefore

*f*_{xx}*= (3x*^{2} *+ 2xy*^{3})

*= 6x + 2y*^{3}

*Example 7 – Solution*

*f*_{xy}*= (3x*^{2} *+ 2xy*^{3})

*= 6xy*^{2}

*f*_{yx}*= (3x*^{2}*y*^{2} *– 4y)*

*= 6xy*^{2}

*f** _{yy}* =

*(3x*

^{2}

*y*

^{2}

*– 4y)*

*= 6x*^{2}*y – 4*

cont’d

33

### Higher Derivatives

*Notice that f*_{xy}*= f** _{yx}* in Example 7. This is not just a

coincidence. It turns out that the mixed partial derivatives
*f*_{xy}*and f** _{yx}* are equal for most functions that one meets in
practice.

The following theorem, which was discovered by the

French mathematician Alexis Clairaut (1713–1765), gives
*conditions under which we can assert that f*_{xy}*= f** _{yx}*.

### Higher Derivatives

Partial derivatives of order 3 or higher can also be defined.

For instance,

and using Clairaut’s Theorem it can be shown that
*f*_{xyy}*= f*_{yxy}*= f** _{yyx}* if these functions are continuous.

35

### Partial Differential Equations

### Partial Differential Equations

*Partial derivatives occur in partial differential equations that *
express certain physical laws.

For instance, the partial differential equation

**is called Laplace’s equation after Pierre Laplace **
(1749–1827).

**Solutions of this equation are called harmonic functions; **

they play a role in problems of heat conduction, fluid flow, and electric potential.

37

### Example 9

*Show that the function u(x, y) = e*^{x}*sin y is a solution of *
Laplace’s equation.

Solution:

We first compute the needed second-order partial derivatives:

*u*_{x}*= e*^{x}*sin y u*_{y}*= e*^{x}*cos y*
*u*_{xx}*= e*^{x}*sin y u*_{yy}*= –e*^{x}*sin y*
*So u*_{xx}*+ u*_{yy}*= e*^{x}*sin y – e*^{x}*sin y = 0*
*Therefore u satisfies Laplace’s equation.*

### Partial Differential Equations

**The wave equation**

describes the motion of a waveform, which could be an ocean wave, a sound wave, a light wave, or a wave

traveling along a vibrating string.

39

### Partial Differential Equations

*For instance, if u(x, t) represents the displacement of a *

*vibrating violin string at time t and at a distance x from one *
*end of the string (as in Figure 8), then u(x, t) satisfies the *
wave equation.

*Here the constant a depends on the density of the string *
and on the tension in the string.

**Figure 8**

### Partial Differential Equations

Partial differential equations involving functions of three variables are also very important in science and

engineering. The three-dimensional Laplace equation is

*and one place it occurs is in geophysics. If u(x, y, z) *

*represents magnetic field strength at position (x, y, z), then *
it satisfies Equation 5. The strength of the magnetic field
indicates the distribution of iron-rich minerals and reflects
different rock types and the location of faults.

41

### Partial Differential Equations

Figure 9 shows a contour map of the earth’s magnetic field as recorded from an aircraft carrying a magnetometer and flying 200 m above the surface of the ground. The contour map is enhanced by color-coding of the regions between the level curves.

Magnetic field strength of the earth

**Figure 9**

### Partial Differential Equations

Figure 10 shows a contour map for the second-order partial
*derivative of u in the vertical direction, that is, u** _{zz}*. It turns

*out that the values of the partial derivatives u*

_{xx}*and u*

*are relatively easily measured from a map of the magnetic field.*

_{yy}*Then values of u** _{zz}* can be calculated from Laplace’s
equation (5).

Second vertical derivative of the magnetic field

**Figure 10**

43

### The Cobb-Douglas Production

### Function

### The Cobb-Douglas Production Function

We have described the work of Cobb and Douglas in

*modeling the total production P of an economic system as *
*a function of the amount of labor L and the capital *

*investment K. *

Here we use partial derivatives to show how the particular form of their model follows from certain assumptions they made about the economy.

*If the production function is denoted by P = P(L, K), then *
the partial derivative *∂P/∂L is the rate at which production *
changes with respect to the amount of labor. Economists
call it the marginal production with respect to labor or the
**marginal productivity of labor.**

45

### The Cobb-Douglas Production Function

Likewise, the partial derivative *∂P/∂K is the rate of change *
of production with respect to capital and is called the

**marginal productivity of capital. **

In these terms, the assumptions made by Cobb and Douglas can be stated as follows.

(i) If either labor or capital vanishes, then so will production.

(ii) The marginal productivity of labor is proportional to the amount of production per unit of labor.

(iii) The marginal productivity of capital is proportional to the amount of production per unit of capital.

### The Cobb-Douglas Production Function

*Because the production per unit of labor is P/L, assumption *
(ii) says that

for some constant α.

*If we keep K constant (K = K*_{0}), then this partial differential
equation becomes an ordinary differential equation:

47

### The Cobb-Douglas Production Function

If we solve this separable differential equation, we get
*P(L, K*_{0}*) = C*_{1}*(K*_{0}*)L*^{α}

*Notice that we have written the constant C*_{1 }as a function of
*K*_{0} *because it could depend on the value of K*_{0}.

Similarly, assumption (iii) says that

and we can solve this differential equation to get
*P(L*_{0}*, K) = C*_{2}*(L*_{0}*)K*^{β}

### The Cobb-Douglas Production Function

Comparing Equations 7 and 8, we have
*P(L, K) = bL*^{α}*K*^{β}

*where b is a constant that is independent of both L and K. *

Assumption (i) shows that α > 0 and β > 0.

Notice from Equation 9 that if labor and capital are both
*increased by a factor m, then *

*P(mL, mK) = b(mL)*^{α}*(mK)*^{β} *= m*^{α}^{+}^{β}*bL*^{α}*K*^{β} *= m*^{α}^{+}^{β}*P(L, K)*

49

### The Cobb-Douglas Production Function

If α + β *= 1, then P(mL, mK) = mP(L, K), which means that *
*production is also increased by a factor of m. That is why *
Cobb and Douglas assumed that α + β = 1 and therefore

*P(L, K) = bL*^{α}*K*^{1 –}^{α}

This is the Cobb-Douglas production function.