References

Differential Calculus of Functions Main Reference

of More Than One Variable L. Leithold, The Calculus 7. HarperCollins College Publishing,
Pennsylvania, USA, 1996.
Calculus with Analytical Geometry 2
Other References

Karlo S. Orge M. H. Protter and C. B. Morrey, College Calculus with Analytic

Geometry. Addison-Wesley Publishing Company, California, USA,
Department of Mathematics and Statistics 1977.
College of Science and Mathematics
MSU-Iligan Institute of Technology
R. Smith and R. Minton, Calculus, Third Edition. McGraw-Hill, New
Andres Bonifacio Avenue, Tibanga, 9200 Iligan City York, NY, 2008.
J. Stewart, Calculus, Seventh Edition. Brooks/Cole, Belmont, CA,
2012.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 1 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 2 / 98

3.1 Functions of More Than One Variable 3.1 Functions of More Than One Variable

Definition
The set of all ordered n-tuples of real numbers is called the n-dimensional Remarks
Euclidean space and is denoted by Rn , that is, 1 The domain of f is contained in Rn and the range of f is contained
Rn = {(x1 , x2 , . . . , xn ) : xi ∈ R for all i}. in R.
2 As usual, we use the notation f : D → R to denote a function f with
An ordered n-tuple (x1 , x2 , . . . , xn ) in Rn is called a point in Rn .
domain D ⊆ Rn and range in R. In this case, if
P = (a1 , a2 , . . . , an ) ∈ D, the value of the function f at P will be
Definition denoted by f (P ) or f (a1 , a2 , . . . , an ).
A function in n variables is a set of ordered pairs of the form (P, w) with
3 If the domain D ⊆ Rn of a function f is not specified, then by
P ∈ Rn and w ∈ R such that no two distinct ordered pairs have the same
default, D is the set of all points P ∈ Rn for which f (P ) is defined
first element. The set of all admissible values of P is called the domain of
(i.e., f (P ) is a real number).
the function and the set of all resulting values of w is called the range of
the function.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 3 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 4 / 98
3.1 Functions of More Than One Variable 3.1 Functions of More Than One Variable
Example 1 Example 1
The following examples show how to determine the domains of certain functions p
2 For the function g(x, y) = x2 − 4y 2 + 16, g is defined when
in two variables:
y2 x2
x−y x2 − 4y 2 + 16 ≥ 0, or equivalently, − ≤ 1. The domain of g consists
1 For the function f defined by f (x, y) = , f is defined when x + y ̸= 0. 4 16
x+y y2 x2
Hence, the domain of f is all of the points in R2 not lying on the line of all points between the two branches of the hyperbola − = 1,
4 16
x + y = 0: including the hyperbola itself:

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 5 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 6 / 98

3.1 Functions of More Than One Variable 3.1 Functions of More Than One Variable
The graph of a function f in two variables, say x and y, is the surface Another way of representing the function f in two variables x and y geometrically
z = f (x, y), where (x, y) is in the domain of f . is through the use of level curves.

Example 2 Definition
x−y Let f be a function in two variables x and y. A level curve of f is the (two-
The graphs of the functions f and g defined by f (x, y) = and dimensional) graph of the equation f (x, y) = c, where c is a constant. In other
p x+y words, a level curve of f is the trace of the surface z = f (x, y) to the plane z = c.
g(x, y) = x2 − 4y 2 + 16 are shown below.

x−y
(a) z = p
Figure: Level curve of a function (shown in blue)
x+y (b) z = x2 − 4y 2 + 16

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 7 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 8 / 98
3.1 Functions of More Than One Variable 3.1 Functions of More Than One Variable
Example 3
The following figure represents some level curves of the function g defined
by p
g(x, y) = x2 − 4y 2 + 16. While the graph for functions in three variables x, y, z requires a
four-dimensional space, which is difficult to visualize in its entirety, we can
represent them geometrically using level surfaces.

Definition
Let f be a function in three variables x, y, and z. A level surface of f is
the (three-dimensional) graph of the equation f (x, y, z) = c, where c is a
constant.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 9 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 10 / 98

3.1 Functions of More Than One Variable 3.2 Partial Derivatives

Recall that for a function f of one variable, say x, the derivative (function)
Example 4 df
of f , denoted by f ′ , Dx f or , is the function whose value at x in the
The following figures represent some level surfaces of the function F dx
domain of f is given by
defined by F (x, y, z) = x2 + y 2 − z 2 at c = −1, 0, 1.
f (x + h) − f (x)
f ′ (x) = lim ,
h→0 h
if this limit exists.
Definition
Let f be a function in two variables x and y. The partial derivative of
f with respect to x, is the function denoted by fx such that its function
value at any point (x, y) in the domain of f is given by
(a) F (x, y, z) = −1 (b) F (x, y, z) = 0 (c) F (x, y, z) = 1
f (x + ∆x, y) − f (x, y)
fx (x, y) = lim ,
∆x→0 ∆x
Figure: Level surfaces of F (x, y, z) = x2 + y 2 − z 2
if this limit exists.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 11 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 12 / 98
3.2 Partial Derivatives 3.2 Partial Derivatives

Remarks
Definition
1 The derivative of a function in a single variable is usually called the
The partial derivative of f with respect to y, is the function denoted
ordinary derivative.
by fy such that its function value at any point (x, y) in the domain of f is
given by 2 Let f be a function in x and y. The following table represents the
f (x, y + ∆y) − f (x, y) notation for partial derivatives of f :
fy (x, y) = lim ,
∆y→0 ∆y
Partial Derivative subscript Leibniz “D”
if this limit exists. of f notation notation operator
Example 5 ∂f
with respect to x fx D1
Use the definition of the partial derivatives to find fx (x, y) and fy (x, y), ∂x
given that f (x, y) = 2x2 − y 2 + 3. ∂f
with respect to y fy D2
∂y

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 13 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 14 / 98

3.2 Partial Derivatives 3.2 Partial Derivatives

Remarks
3 In finding the partial derivative of f with respect to x, we Definition
differentiate the function treating y as a constant.
Let f be a function in n variables x1 , x2 , . . . , xn . The partial derivative of
4 In finding the partial derivative of f with respect to y, we differentiate f with respect to xk , is the function denoted by fxk such that its function
the function treating x as a constant. value at any point P (x1 , x2 , . . . , xn ) in the domain of f is given by
5 The theorems on differentiation are applicable to partial derivatives.
fxk (x1 , x2 , . . . , xn )
Example 6 f (x1 , x2 , . . . , xk + ∆xk , . . . , xn ) − f (x1 , x2 , . . . , xn )
= lim ,
∂f ∂f ∆xk →0 ∆xk
Given the following functions, find and :
∂x ∂y
if this limit exists.
1 f (x, y) = 2x3 y 2 − 3x2 y 3 + 6y 2 − 2x3 + 10
2 f (x, y) = yex sin(3y) + e−y cos(2x)

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 15 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 16 / 98
3.2 Partial Derivatives 3.3 Higher-Order Partial Derivatives
∂f ∂f
If f is a function in x and y, then and are functions of x and y
∂x ∂y
also. Hence, we may define the following:
Remarks Second-Order Partial Derivatives
When taking the partial derivative of a function with respect to one variable, ∂2f
 
∂ ∂f ∂f
the other variables will be treated as constants. 1
2
= = fxx is the partial derivative of with respect
∂x ∂x ∂x ∂x
to x
Example 7 ∂2f
 
∂ ∂f ∂f
Find the indicated partial derivative of the following functions: 2 = = fyy is the partial derivative of with respect
∂y 2 ∂y ∂y ∂y
2 2 2 ∂w to y
1 w = x y + y z + z x;
∂2f
 
∂x ∂ ∂f ∂f
2 r 3 = = fxy is the partial derivative of with respect
2 f (r, θ, ϕ) = 4r sin θ + 5e cos θ sin ϕ − 2 cos ϕ; fθ (r, θ, ϕ) ∂y∂x ∂y ∂x ∂x
to y
∂2f
 
∂ ∂f ∂f
4 = = fyx is the partial derivative of with respect
∂x∂y ∂x ∂y ∂y
to x

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 17 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 18 / 98

3.3 Higher-Order Partial Derivatives 3.3 Higher-Order Partial Derivatives

Remarks
1 The other higher-order derivatives are defined similarly.
Third-Order Partial Derivatives
2 There are various notations for a specific derivative. For example, if f
∂3f ∂2f
  
∂ ∂ ∂f
1 = = fxxx is the partial derivative of with is a function in two variables x and y, then
∂x3 ∂x ∂x ∂x ∂x2
respect to x
∂3f ∂3f
∂3f ∂2f fxxy D112
  
∂ ∂ ∂f ∂y∂x∂x ∂y∂x2
2 = = fxyx is the partial derivative of
∂x∂y∂x ∂x ∂y ∂x ∂y∂x
with respect to x all stand for the third partial derivative of f that is obtained by
∂3f ∂2f partial-differentiating twice with respect to x and then once with
  
∂ ∂ ∂f
3 = = fyyx is the partial derivative of respect to y.
∂x∂y∂y ∂x ∂y ∂y ∂y 2
with respect to y 3 In the subscript notation and the “D” operator notation, the order of
and so on. partial differentiation is from left to right.
4 In the Leibniz notations, the order of partial differentiation is from
right to left.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 19 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 20 / 98
3.3 Higher-Order Partial Derivatives 3.3 Higher-Order Partial Derivatives

Theorem
Example 8 If fxy (x, y) and fyx (x, y) are continuous on an open disk containing (x0 , y0 ),
Find the indicated partial derivative. then fxy (x0 , y0 ) = fyx (x0 , y0 ).
1 f (x, y) = 2x3 y 2 − 3x2 y 3 ; fxx (x, y)
∂3F The following example shows that there is a function f where
2 F (x, y, z) = xy 2 cos(ez ); fxy (x0 , y0 ) ̸= fyx (x0 , y0 ).
∂x∂y 2
Example 10
Example 9
Suppose that we are going to compute fxy (0, 0) and fyx (0, 0) for the
Given that
function f defined by
f (x, y) = x3 y − y cosh(xy),
 xy(x2 − y 2 )

find fxy (x, y) and fyx (x, y). , (x, y) ̸= (0, 0);
f (x, y) = x2 + y 2
0, (x, y) = (0, 0).


Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 21 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 22 / 98

3.3 Higher-Order Partial Derivatives 3.3 Higher-Order Partial Derivatives

Example 10 Example 10
First, we compute the following: if y ̸= 0, Also, if x ̸= 0,

f (0 + ∆x, y) − f (0, y) f (x, 0 + ∆y) − f (x, 0)

fx (0, y) = lim fy (x, 0) = lim
∆x→0 ∆x ∆y→0 ∆y



x(∆y) x − (∆y)2
2
x(0) x2 − 02



∆xy (∆x)2 − y 2 0y 02 − y 2
− −
(∆x)2 + y 2 02 + y 2 x2 + (∆y)2 x2 + 0 2
= lim = lim
∆y

∆x
∆x→0 ∆y→0
y (∆x)2 − y 2 y(02 − y 2 )



x x2 − (∆y)2 x x2 − 02
= lim = = −y. = lim = = x,
∆x→0 (∆x)2 + y 2 02 + y 2 ∆y→0 x2 + (∆y)2 x2 + 02
and and


∆x(0) (∆x)2 − 02


0(∆y) 02 − (∆y)2
−0 −0
f (0 + ∆x, 0) − f (0, 0) (∆x)2 + 02 f (0, 0 + ∆y) − f (0, 0) 02 + (∆y)2
fx (0, 0) = lim = lim = 0. fy (0, 0) = lim = lim = 0.
∆x→0 ∆x ∆x→0 ∆x ∆y→0 ∆y ∆y→0 ∆y

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 23 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 24 / 98
3.3 Higher-Order Partial Derivatives 3.3 Higher-Order Partial Derivatives

Example 10
Now,
Example 11
fx (0, 0 + ∆y) − fx (0, 0) −∆y − 0
fxy (0, 0) = lim = lim = −1, Show that u = (x2 + y 2 + z 2 )−1/2 satisfies the Laplace’s equation in R3 :
∆y→0 ∆y ∆y→0 ∆y

but ∂2u ∂2u ∂2u

+ 2 + 2 = 0.
∂x2 ∂y ∂z
fy (0 + ∆x, 0) − fy (0, 0) ∆x − 0
fyx (0, 0) = lim = lim = 1.
∆x→0 ∆x ∆x→0 ∆x
Thus, fxy (0, 0) ̸= fyx (0, 0).

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 25 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 26 / 98

3.3 Higher-Order Partial Derivatives 3.4 The Total Differential

Exercises 12
x y
1 Find zx and zy given that z = sin
y
+ ln .
x
Definition
If f is a function in two variables x and y, and f is differentiable at (x, y),
∂u z2
2 Find given that u = p . then the total differential of f is the function df having values given by
∂z x2 + y 2 + z 2
∂f (x, y) ∂f (x, y)
3 Let f (x, y) = ex cos y + 2x2 y + ln(xy 2 ). Show that fyxx − fyx + 4x = 4. df (x, y, ∆x, ∆y) = ∆x + ∆y.
∂x ∂y
4 Show that the function u(x, y, z) = e3x+4y sin(5z) satisfies the Laplace’s
equation in R3 : Remarks
∂2u ∂2u ∂2u
+ 2 + 2 = 0. If we let z = f (x, y), we usually write dz in place of df (x, y, ∆x, ∆y). Thus,
∂x2 ∂y ∂z
if we let dx = ∆x and dy = ∆y, then the total differential can be written
5 Show that u(t, x) = Ae−4αt sin(2x) satisfies the one-dimensional heat as ∂z ∂z
equation dz = dx + dy.
∂x ∂y
∂u ∂2u
−α 2 =0
∂t ∂x
for any constants A and α.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 27 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 28 / 98
3.4 The Total Differential 3.4 The Total Differential

One application of the total differential is to estimate the change of a

dependent variable given some changes on the independent variables.
Definition Example 14
If f is a function in n variables x1 , x2 , . . . , xn and f is differentiable at 1 The power consumed in an electrical resistor is given by
(x1 , x2 , . . . , xn ), then the total differential of f is given by E2
∂f ∂f ∂f P = (in watts).
df = dx1 + dx2 + · · · + + dxn . R
∂x1 ∂x2 ∂xn If E = 200 volts and R = 8 ohms, approximately how much does the
power change if E is decreased by 5 volts and R is decreased by 0.2
Example 13 ohm?
Find the total differential dw if w = exy − sin(yz). 2 A closed container in the shape of a rectangular solid is to have an
inside length of 8 m, an inside width of 5 m, an inside height of 4 m,
and a thickness of 4 cm. Find the approximate amount of material
needed to construct the container.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 29 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 30 / 98

3.4 The Total Differential 3.4 The Total Differential

Example 15
Another application of the total differential is approximating the largest 1 Find approximately the greatest error in calculating the hypotenuse of
possible error of computing the dependent variable given some errors in a right triangle from the lengths of the legs if they are measured to be
measuring the independent variables. For instance, if z = f (x, y), the 6 cm and 8 cm, respectively, with the possible error of 0.1 cm for
largest possible error of computing z is the maximum value of |dz|. From each measurement. Find the approximate percent error.
the total differential 2 The specific gravity s of an object is given by the formula
∂z ∂z
dz = dx + dy,
∂x ∂y A
s=
using the properties of absolute value, we obtain A−W
where A is the number of pounds in the weight of the object in air
∂z ∂z ∂z ∂z ∂z ∂z
|dz| = dx + dy ≤ dx + dy = |dx| + |dy|. and W is the number of pounds in the weight of the object in water.
∂x ∂y ∂x ∂y ∂x ∂y
If the weight of an object in air is read 20 lb with a possible error of
0.01 lb and its weight in water is read as 12 lb with a possible error of
0.02 lb, find approximately the largest error in calculating s from
these measurements. Find the approximate relative error.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 31 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 32 / 98
3.4 The Total Differential 3.4 The Total Differential

Exercises 16 Exercises 16
1 The production function P for one country is 3 A housing lot is in the shape of a trapezoid. The measurement of the
bases are found to be 38 m and 42 m, respectively, while the
P (l, c) = l0.65 c0.35 , measurement of the height is 30 m. If all measurements have a
possible error of 5 cm, find approximately the greatest error in
where l stands for units of labor and c for the units of capital. At calculating the area of the housing lot.
present, 50 units of labor and 29 units of capital are available. Use the
4 Suppose that α is the radian measure of an acute angle of a right
total differentials to estimate the change in production as the number
triangle and sin α is determined by ac , where a cm is the length of the
of units is increased to 52 and capital is decreased to 27 units.
leg opposite α, and c cm is the length of the hypotenuse. If by
2 A conical tank is to be made with radius of base r = 2 m and height measurement a is found to be 3.52 and c is found to be 7.14, and
h = 5 m. If r is increased by 0.2 m and h is decreased by 0.3 m, there is a possible error of 0.01 in each measurement, find
approximate using the total differential the change in volume of the approximately the greatest possible error in the computation of sin α
tank. from these measurements.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 33 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 34 / 98

3.5 Tangent Plane Approximation of Functions in 3.5 Tangent Plane Approximation of Functions in
Two Variables Two Variables

Recall that given a function f of a single variable x which is differentiable

at x = a, the slope of the tangent line to the curve y = f (x) at x = a is Theorem
equal to f ′ (a), so that an equation of the said tangent line is given by
Suppose that f (x, y) has continuous first partial derivatives at (a, b). A
⇀
′
y = f (a) + f (a)(x − a). normal vector N to the tangent plane to z = f (x, y) at (a, b) is
⇀
We also call this the linear approximation to f (x) at x = a. N = ⟨fx (a, b), fy (a, b), −1⟩.

Moreover, an equation of this tangent plane is given by

We may also approximate the value of a function of two variables near a
given point using the tangent plane to the surface at that point. Our main fx (a, b)(x − a) + fy (a, b)(y − b) − (z − f (a, b)) = 0.
task is to look for an equation of the tangent plane to z = f (x, y) at the
point (a, b, f (a, b)), where fx and fy are continuous at (a, b).

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 35 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 36 / 98
3.5 Tangent Plane Approximation of Functions in 3.6 Chain Rule
Two Variables

Theorem (The Chain Rule)

Example 17 If u is a differentiable function in x and y, defined by u = f (x, y), where
∂x ∂x ∂y ∂y
Find an equation of the tangent plane to the surface x = F (r, s), y = G(r, s) and , , , and all exist, then u is a
∂r ∂s ∂r ∂s
function in r and s,
x2
z = x3 + y 3 +
y ∂u ∂u ∂x ∂u ∂y ∂u ∂u ∂x ∂u ∂y
= · + · and = · + ·
∂r ∂x ∂r ∂y ∂r ∂s ∂x ∂s ∂y ∂s
at the point (2, 1, 13).
Remarks
In future sections, we will calculate equations of tangent planes and other In the above theorem, u is the dependent variable, r and s are independent
related geometric objects for general surfaces. variables, and x and y are called the intermediate variables.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 37 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 38 / 98

3.6 Chain Rule 3.6 Chain Rule

Theorem (The Generalized Chain Rule)

If u is a differentiable function in n variables x1 , x2 , . . . , xn , and each of
these variables is in turn a function in m variables y1 , y2 , . . . , ym . Suppose
∂xi Example 18
further that each of the partial derivatives (i = 1, 2, . . . , n, j =
∂yj ∂u
1 Find if u = ey/x ; x = 2r cos s; and y = 4r sin s.
1, 2, . . . , m) exists. Then u is a function in y1 , y2 , . . . , ym , and ∂r
2 Let f be a function in x and y, where x = r cos θ and y = r sin θ.
∂u ∂u ∂x1 ∂u ∂x2 ∂u ∂xn Show that
= · + · + ··· + ·
∂y1 ∂x1 ∂y1 ∂x2 ∂y1 ∂xn ∂y1 fθ = −fx r sin θ + fy r cos θ.
∂u ∂u ∂x1 ∂u ∂x2 ∂u ∂xn
= · + · + ··· + ·
∂y2 ∂x1 ∂y2 ∂x2 ∂y2 ∂xn ∂y2
..
.
∂u ∂u ∂x1 ∂u ∂x2 ∂u ∂xn
= · + · + ··· + ·
∂ym ∂x1 ∂ym ∂x2 ∂ym ∂xn ∂ym

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 39 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 40 / 98
3.6 Chain Rule 3.6 Chain Rule (Implicit Partial Differentiation)
Partial differentiation can be done for functions defined implicitly.
Exercises 19
Example 20
∂w
1 Find if w = xy + yz + xz; x = r + s + t; y = rst; and ∂z
∂t Assume that z is a function in x and y. Find if
z = r 2 + s2 + t2 . ∂x
2 Let f be a function in x and y, where x = r cos θ and y = r sin θ. 3x2 + y 2 + z 2 = 3xy − 4xz + 15.
Show that
fr = fx cos θ + fy sin θ One of the applications of Chain Rule for partial derivatives is that it
produces a convenient formula for implicit partial differentiation.
and
frr = fxx cos2 θ + 2fxy cos θ sin θ + fyy sin2 θ. Theorem
3 Suppose that z = f (x, y), x = uv, and y = u2 + v 2 . If fx (1, 2) = 2, If f is a differentiable function in x and y such that z = f (x, y) and f is
fy (1, 2) = 3, fxy (1, 2) = 1, fyx (1, 2) = 1, fxx (1, 2) = 0 and defined implicitly by the equation F (x, y, z) = 0, then
fyy (1, 2) = −2, calculate zu and zuv at the point where u = 1 and Fy (x, y, z)
∂z Fx (x, y, z) ∂z
v = 1. =− and =−
∂x Fz (x, y, z) ∂y Fz (x, y, z)

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 41 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 42 / 98

3.6 Chain Rule (Implicit Partial Differentiation) 3.6 Chain Rule (Implicit Partial Differentiation)

The Chain Rule for partial derivatives is also used to obtain a convenient
formula for implicit ordinary differentiation.
Example 21 Theorem
Assume that z is a function in x and y. Use the previous theorem to find If y is a differentiable function in x, and is defined implicitly by the equation
the indicated partial derivative. F (x, y) = 0, then
2 2 2 ∂z dy Fx (x, y)
1 3x + y + z = 3xy − 4xz + 15; =−
∂x dx Fy (x, y)
2 2 ∂z
2 z = (x + y ) sin xz;
∂y Example 22
Assuming that y is a differentiable function in x, find y ′ if

x3 + y 3 = 6xy.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 43 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 44 / 98
3.6 Chain Rule (Implicit Partial Differentiation) 3.6 Chain Rule (Total Derivative and Related Rates)
Suppose that u is a differentiable function in x and y, and both x and
Exercises 23 y are differentiable functions in the single variable t. Recall that by chain
1 Assume that z is a function in x and y. Find the indicated partial rule,
derivative. ∂u ∂u ∂x ∂u ∂y
∂z = · + ·
(a) 3x2 z + 2z 3 − 3yz = 0; ∂t ∂x ∂t ∂y ∂t
∂x
∂z
(b) yexyz cos 3xz = 5; Because x and y are differentiable functions of the single variable t, u is
∂y
(c) xyz − 4y 2 z 2 + sin xy = 0; zx also a function in the single variable t. Consequently,
2 Assume that z is a function in x and y. Find zx and zy if ∂u du ∂x dx ∂y dy
= , = , and = .
∂t dt ∂t dt ∂t dt
cos(xyz) = eyz − y + xz. Therefore,
z
3 Consider the equation z = e (sin x + cos y), where z is a function in du ∂u dx ∂u dy
x and y. Show that = · + · .
dt ∂x dt ∂y dt
zx sin y
· = −1. du
zy cos x The ordinary derivative given by the above formula is called the total
dt
derivative of u with respect to t.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 45 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 46 / 98

3.6 Chain Rule (Total Derivative and Related Rates) 3.6 Chain Rule (Total Derivative and Related Rates)

We apply the total derivative in problems involving related rates.

Example 24
Example 24
2 Suppose that the portion of a tree that is usable for lumber is a right
1 According to the ideal gas law for a confined gas, if P atmospheres is circular cylinder. If the usable height of the tree increases 2 feet per
the pressure, V liters is the volume, and T Kelvin is the absolute year and the usable radius increases 18 feet per year, how fast is the
temperature, we have volume of the usable lumber changing at the instant the usable height
P V = kT is 20 feet and the usable radius is 1 41 feet?
where k is a constant of proportionality. Suppose that a quantity of 3 At a given instant, the length of one leg of a right triangle is 10 cm
gas obeys the ideal gas law with k = 1.2, and the gas is in the and it is increasing at a rate of 1 cm/min, and the length of the other
container that is heated at the rate of 3 K/min. If at the instant leg is 12 cm and it is decreasing at a rate of 2 cm/min. Find the rate
when the temperature is 300 K, the pressure is 6 atm and is of change of the measure of the acute angle opposite the leg of
decreasing at the rate of 0.1 atm/min, find the rate of change of the length 12 cm at the given instant.
volume at that instant.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 47 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 48 / 98
3.6 Chain Rule (Total Derivative and Related Rates) 3.6 Chain Rule (Total Derivative and Related Rates)

Exercises 25
1 The volume of a right circular cylinder is given by the formula
V = πr2 h. Suppose that its radius is decreasing at the rate of
2.5 cm/min and its height is increasing at the rate of 6 cm/min. Find
the rate of change of the volume at the instant when the radius is
Exercises 25
10 cm and the height is 20 cm. 4 A light shines from the top of a tower 10 m high. A ball is dropped
from the same height but 5 m away from the tower, and its height
2 In △ABC, the length of AC is increasing at a rate of 2 cm/min,
from the ground at time t seconds is given by h(t) = 10 − 12 t2 . How
while the length of BC is decreasing at a rate of 1 cm/min. If ∠C
fast is the ball’s shadow moving on the ground 2 seconds later?
maintains its angle measure of 2π
3 radians, how fast is the length of
side AB changing at the instant AC = 6 cm and BC = 3 cm?
3 The height of a right-circular cone is increasing at the rate of 40
cm/min and the radius is decreasing at the rate of 15 cm/min. Find
the rate of change of the volume at the instant when the height is
200 cm and the radius is 60 cm.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 49 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 50 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
Definition Definition
⇀ ⇀ ⇀ ⇀
A vector A in Rn is an ordered n-tuple A = ⟨a1 , a2 , . . . , an ⟩ of real Let A = ⟨a1 , a2 , . . . , an ⟩ and B = ⟨b1 , b2 , . . . , bn ⟩ be vectors in Rn , and c be a
numbers. The numbers a1 , a2 , . . . , an are called the components of scalar.
⇀ ⇀ ⇀ ⇀ ⇀ ⇀
the vector A , and the number n as the dimension of A . Two vectors 1 The sum of A and B , denoted by A + B , is a vector given by
⟨a1 , a2 , . . . , an ⟩ and ⟨b1 , b2 , . . . , bn ⟩ are equal if and only if ai = bi , for all ⇀ ⇀
A + B = ⟨a1 + b1 , a2 + b2 , . . . , an + bn ⟩.
i = 1, 2, . . . , n. We denote by Vn the set of all n-dimensional vectors. ⇀ ⇀
2 The negative of A , denoted by − A , is a vector given by
⇀
Definition − A = ⟨−a1 , −a2 , . . . , −an ⟩.
1 Real numbers are referred to as scalars. ⇀ ⇀ ⇀ ⇀
⇀ 3 The difference of A and B , denoted by A − B , is a vector given by
2 The vector ⟨0, 0, . . . , 0⟩ is called the zero vector, denoted by 0 . ⇀ ⇀ ⇀ ⇀
| {z } A − B = A + (−B ) = ⟨a1 − b1 , a2 − b2 , . . . , an − bn ⟩.
n zeroes ⇀ ⇀
⇀ ⇀ 4 The scalar product of c and A , denoted by c A , is a vector given by
3 The magnitude of a vector A = ⟨a1 , a2 , . . . , an ⟩, denoted by ∥ A ∥, ⇀
is defined by c A = ⟨ca1 , ca2 , . . . , can ⟩.
⇀ ⇀ ⇀ ⇀ ⇀
q
∥ A ∥ = a21 + a22 + · · · + a2n . 5 The dot product of A and B , denoted by A · B , is a scalar given by
⇀ ⇀
A · B = a1 b1 + a2 b2 + · · · + an bn .
4 A unit vector is a vector of magnitude equal to one.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 51 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 52 / 98
3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
Remarks
Definition ⇀
1 If U is a unit vector in the same direction as a nonzero vector
⇀ ⇀ ⇀
Let A and B be vectors in Rn . A = ⟨a1 , a2 , . . . , an ⟩, then
⇀ ⇀ ⇀ ⇀ * +
1 A and B are parallel if A = cB for some scalar c. ⇀ 1 ⇀ a1 a2 an
⇀ ⇀ ⇀ ⇀ U = ⇀ A = ⇀ , ⇀ ,..., ⇀ .
2 A and B are orthogonal or perpendicular if A · B = 0. ∥A∥ ∥A∥ ∥A∥ ∥A∥
⇀ ⇀ ⇀ ⇀
3 If A and B are nonzero vectors, then the angle between A and B ⇀ ⇀
2 The definition of the angle between nonzero vectors A and B stems
is the angle θ whose measure is 2 3
from the alternative formula of the dot product in R or R :
⇀ ⇀ ! ⇀ ⇀ ⇀ ⇀
A·B A · B = ∥ A ∥∥B ∥ cos θ.
θ = cos−1 ⇀ ⇀ .
∥ A ∥∥B ∥ 3 Any vector is parallel and orthogonal to the zero vector of the same
⇀ ⇀ dimension.
4 Nonzero vectors A and B are in the same direction if the angle ⇀ ⇀
4 Nonzero vectors A and B are parallel if and only if their angle
between them is 0 radians, while they are of opposite directions if between them is either 0 or π radians.
the angle between them is π radians. ⇀ ⇀
5 Nonzero vectors A and B are orthogonal if and only if their angle
between them is π/2 radians.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 53 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 54 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
For geometric representations, we usually use two or three-dimensional vectors, From the definition of the partial derivative of a function f of n
so there are additional definitions exclusive for them:
variables x1 , x2 , . . . , xn with respect to the k th variable xk at the point
Vector two-dimensional three-dimensional A(a1 , a2 , . . . , an ):
unit basis ı̂ = ⟨1, 0⟩, ȷ̂ = ⟨0, 1⟩ ı̂ = ⟨1, 0, 0⟩, ȷ̂ = ⟨0, 1, 0⟩, f (a1 , a2 , . . . , ak + h, . . . , an ) − f (A)
vectors k̂ = ⟨0, 0, 1⟩ Dk f (A) = lim ,
h→0 h
⇀ ⇀
vector A = ⟨a1 , a2 ⟩ = a1 ı̂ + a2 ȷ̂ A = ⟨a1 , a2 , a3 ⟩
one can see its physical interpretation:
notation
= a1 ı̂ + a2 ȷ̂ + a3 k̂
direction of direction angle: the angle θ direction angles: the Dk f (A) represents the instantaneous rate of change, at A, of the
a nonzero measured from the positive x- nonnegative angles α, β and quantity f per unit change in xk (xk varies alone, and the rest of the
vector axis γ measured from the positive variables are held fixed at xi = ai , i ̸= k).
x, y and z-axis, respectively
⇀ ⇀ In particular, for a function f of two variables x and y, fx (x0 , y0 ) describes
unit vector U = ⟨cos θ, sin θ⟩ U = ⟨cos α, cos β, cos γ⟩ the rate of change, at (x0 , y0 ), of f in the direction of the x-axis (y = y0
⇀
U in a given
is fixed); while fy (x0 , y0 ) describes the rate of change, at (x0 , y0 ), of f in
direction
the direction of the y-axis (x = x0 is fixed).

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 55 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 56 / 98
3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
A geometrical interpretation of the partial derivative is also available The partial derivative fy (x0 , y0 ) is the slope of the tangent line to the
for functions of two variables x and y. The partial derivative fx (x0 , y0 ) is trace of the surface z = f (x, y) with the plane x = x0 at the point (x0 , y0 ).
the slope of the tangent line to the trace of the surface z = f (x, y) with
the plane y = y0 at the point (x0 , y0 ).

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 57 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 58 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
Remarks
We now generalize the definition of a partial derivative to obtain the rate
1 For a function f of two variables x and y and unit vector
of change of a function with respect to any direction. ⇀
U = ⟨u1 , u2 ⟩, the definition of D⇀f (x, y) becomes
U
Definition
⇀ f (x + u1 h, y + u2 h) − f (x, y)
Let f be a function in n variables x1 , x2 , . . . , xn . If U is the unit vector D⇀f (x, y) = lim .
U h→0 h
⟨u1 , u2 , . . . , un ⟩, then the directional derivative of f in the direction of
⇀ ⇀
U , denoted by D⇀f , is given by 2 If U = ı̂, then u1 = 1 and u2 = 0, so that
U
D⇀f (x1 , x2 , . . . , xn ) f (x + h, y) − f (x, y)
U Dı̂ f (x, y) = lim = fx (x, y).
h→0 h
f (x1 + u1 h, x2 + u2 h, . . . , xn + un h) − f (x1 , x2 , . . . , xn )
= lim ,
h→0 h In other words, the partial derivative of f with respect to x is the
directional derivative of f in the direction of ı̂. By similar arguments,
if this limit exists.
the partial derivative of f with respect to y is the directional
derivative of f in the direction of ȷ̂.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 59 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 60 / 98
3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients

Remarks
3 For a function f of three variables x, y and z, and unit vector
⇀
U = ⟨u1 , u2 , u3 ⟩,

f (x + u1 h, y + u2 h, z + u3 h) − f (x, y, z)
D⇀f (x, y, z) = lim .
U h→0 h
4 The partial derivatives fx , fy and fz are the directional derivatives of
f in the direction of ı̂, ȷ̂ and k̂, respectively.

Example 26
Find D⇀f (x, y) if
U
Figure: The directional derivative D⇀f (x0 , y0 ) is the slope of the tangent line to the
f (x, y) = 12 − x2 − 4y 2
U ⇀ π
trace of the surface z = f (x, y) with the plane P perpendicular to the xy-plane which
⇀
and U is the unit vector whose direction angle is 6 radians.
contains P0 (x0 , y0 , f (x0 , y0 )) and the vector U regarded as a three-dimensional vector
drawn in the xy-plane with initial point (x0 , y0 ).
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 61 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 62 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients

Theorem
⇀
If f is a differentiable function of n variables, U = ⟨u1 , u2 , . . . , un ⟩ is a unit
vector, and P is in the domain of f , then Definition
If f is a function of n variables such that all first-order partial derivatives
D⇀f (P ) = u1 [D1 f (P )] + u2 [D2 f (P )] + · · · + un [Dn f (P )]. exist, and P is a point in the domain of f , then the gradient of f at P ,
U
denoted by ∇f (P ) (read as “del f of P ”), is defined by
Remarks
∇f (P ) = ⟨D1 f (P ), D2 f (P ), . . . , Dn f (P )⟩.
From the previous result, the directional derivative D⇀f (P ) can be
U
expressed as the dot product of two vectors:
D⇀f (P ) = ⟨u1 , u2 , . . . , un ⟩ · ⟨D1 f (P ), D2 f (P ), . . . , Dn f (P )⟩.
U

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 63 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 64 / 98
3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
Remarks
1 For a function f of two variables x and y,
Let f be a differentiable function of n variables, and P a point in the
∇f (x, y) = ⟨fx (x, y), fy (x, y)⟩. ⇀ ⇀
domain of f . If ∇f (P ) = 0 , then for all unit vectors U in Rn ,
2 For a function f of three variables x, y and z,
⇀
∇f (x, y, z) = ⟨fx (x, y, z), fy (x, y, z), fz (x, y, z)⟩. D⇀(P ) = U · ∇f (P ) = 0.
U
⇀
3 If a function f of n points is differentiable, U is an n-dimensional unit In this case, there is no instantaneous rate of change at P .
vector, and P is a point in the domain of f , then
⇀ ⇀
D⇀f (P ) = U · ∇f (P ). If ∇f (P ) ̸= 0 , and θ is the radian measure of the angle between the
U ⇀
two vectors U and ∇f (P ),
⇀ ⇀
Example 27 D⇀f (P ) = U · ∇f (P ) = ∥ U ∥∥∇f (P )∥ cos θ = ∥∇f (P )∥ cos θ.
U
1 Find the gradient of the function f defined by f (x, y, z) = cos(xy) + z at
P (0, −2, 4).
2 Use ➊ to compute the directional derivative of f at P in the direction of the
vector 3ȷ̂ − 4k̂.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 65 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 66 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
If f is a differentiable function of n variables, and P is a point in the
⇀
domain of f such that ∇f (P ) ̸= 0 , then at the point P ,
Now,
1 the direction of steepest ascent (that is, the direction of maximum
1 The maximum value of D⇀f (P ) occurs when cos θ = 1, that is,
⇀ U (instantaneous) rate of change) is in the direction of ∇f (P ), with
when θ = 0, or U is in the same direction as ∇f (P ). maximum slope ∥∇f (P )∥;
2 The maximum value of D⇀f (P ) is ∥∇f (P )∥. 2 the direction of steepest descent (that is, the direction of minimum
U
3 The minimum value of D⇀f (P ) occurs when cos θ = −1, that is, rate of change) is opposite the direction of ∇f (P ), with minimum
⇀ U slope −∥∇f (P )∥; and
when θ = π, or U is in the opposite direction as ∇f (P ).
3 the direction where the rate of change is zero is the direction of any
4 The minimum value of D⇀f (P ) is −∥∇f (P )∥.
U vector orthogonal to ∇f (P ).
⇀
5 D⇀f (P ) = 0 when cos θ = 0, that is, when θ = π2 , or U and ∇f (P )
U Furthermore, for a function of two variables, if P = (x0 , y0 ) lies on a level
are orthogonal. ⇀
curve of f , and D⇀f (x0 , y0 ) = 0, then U is tangent to that level curve at
U
(x0 , y0 ), since the level curves of f are curves on the xy-plane where f is
constant.

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 67 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 68 / 98
3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
The gradient and the directional derivative are used in problems involving
(instantaneous) rate of change per unit length/distance.

Example 28 Example 28
1 An equation of the surface of a mountain is 2 The temperature T (in ◦ C) at any point (x, y, z) in a certain room is
given by
z = 1200 − 3x2 − 2y 2 xyz
T (x, y, z) = .
1 + x + y2 + z2
2
where distance is measured in meters, the x-axis points to the east,
and the y-axis points to the north. A mountain climber is at the (a) Find the rate of change of the temperature at the point (1, 1, 1) in the
point corresponding to (−10, 5, 850). direction toward the point (3, 0, −1).
(a) What is the direction of steepest ascent? (b) A fly at point (1, 1, 1) wants to travel in the direction in which the
(b) If the climber moves in the east direction, is she ascending or temperature drops most rapidly. Find a unit vector in that direction.
descending, and what is her rate?
(c) If the climber moves in the southwest direction, is she ascending or
descending, and what is her rate?
(d) In what direction is she travelling a level path?

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 69 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 70 / 98

3.7 Directional Derivatives and Gradients 3.7 Directional Derivatives and Gradients
Exercises 29
1 Suppose V (x, y, z) volts is the electric potential at any point (x, y, z) in
three-dimensional space and Exercises 29
V (x, y, z) = ze −3x
cos 2y. 3 The temperature T (in ◦ C) at any point (x, y, z) is given by

(a) Find the rate of change of V at the point (0, π3 , 1) in the direction of T (x, y, z) = 200e−x
2 −3y 2 −9z 2
,
the vector ı̂ − ȷ̂ − k̂.
(b) Find the unit vector such that its direction will give the greatest rate of where x, y and z are measured in meters.
change of V at (0, π3 , 1). (a) Find the rate of change of the temperature at the point P (2, −1, 2) in
2 An equation of the surface of the mountain is z = 900 − 3xy, where distance the direction toward the point (3, −3, 3).
is measured in meters, the x-axis points to the west, and the y-axis points to (b) Determine the unit vector in which direction the temperature increases
the south. A mountain climber is at the point corresponding to (50, 4, 300). fastest at P .
(c) Find the fastest rate of increase at P .
(a) What is the direction of steepest ascent?
(b) Is the climber ascending or descending when moving in the north
direction?
(c) In what direction is the climber travelling a level path?

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 71 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 72 / 98
3.8 Tangent Planes to Level Surfaces and Related Concepts 3.8 Tangent Planes to Level Surfaces and Related Concepts
Definition Remarks
A nonzero vector orthogonal to a tangent vector of every curve C through A normal vector to a surface is not unique. In fact, any two normal vectors to a
a point P0 on the surface S is called a normal vector to S at P0 . surface are parallel.

Theorem
If an equation of a surface S is F (x, y, z) = 0, and F is differentiable and Fx , Fy ,
and Fz are not all zero at the point P0 (x0 , y0 , z0 ) on S, then ∇F (x0 , y0 , z0 ) is a
normal vector to S at P0 .

Definition
1 If an equation of a surface S is F (x, y, z) = 0, and F is differentiable and
Fx , Fy , and Fz are not all zero at the point P0 (x0 , y0 , z0 ) on S, then the
tangent plane of S at a point P0 is the plane through P0 having
∇F (x0 , y0 , z0 ) as a normal vector.
⇀ 2 The normal line to a surface S at a point P0 on S is the line through P0 on
Figure: A normal vector N of a surface S at P0 , which is orthogonal to a tangent
⇀ S is the line through P0 having as the set of direction numbers the
vector T of every curve C on S passing through P0
components of any normal vector to S at P0 .
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 73 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 74 / 98

3.8 Tangent Planes to Level Surfaces and Related Concepts 3.8 Tangent Planes to Level Surfaces and Related Concepts

Remarks
1 Since ∇F (x0 , y0 , z0 ) = ⟨Fx (x0 , y0 , z0 ), Fy (x0 , y0 , z0 ), Fz (x0 , y0 , z0 )⟩ ,
an equation of the tangent plane to the surface S at P0 is
Fx (x0 , y0 , z0 )(x−x0 )+Fy (x0 , y0 , z0 )(y−y0 )+Fz (x0 , y0 , z0 )(z−z0 ) = 0,
while a set of parametric equations of the normal line to S at P0 is
x = x0 + Fx (x0 , y0 , z0 )t
y = y0 + Fy (x0 , y0 , z0 )t
z = z0 + Fx (x0 , y0 , z0 )t.

2 If none of the components of of ∇F (x0 , y0 , z0 ) are equal to zero, a

set of symmetric equations of the normal line to S at P0 is
x − x0 y − y0 z − z0
= = .
Figure: The tangent plane and the normal line to a surface at a point Fx (x0 , y0 , z0 ) Fy (x0 , y0 , z0 ) Fz (x0 , y0 , z0 )

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 75 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 76 / 98
3.8 Tangent Planes to Level Surfaces and Related Concepts 3.8 Tangent Planes to Level Surfaces and Related Concepts
Example 30 Consider a curve C which is the intersection of two surfaces having
Find an equation of the tangent plane and a set of symmetric equations of equations F (x, y, z) = 0 and G(x, y, z) = 0.
the normal line to the given surface at the indicated point.
1 x2 + y 2 + z 2 = 17; (2, −2, 3)
x
2 y = e cos z; (1, e, 0)

Figure: The tangent line to the curve of intersection of two surfaces

(a) Item #1 (b) Item #2

We are going to find equations of the tangent line to the curve C at a point
Figure: Illustrations for Example 30
P0 (x0 , y0 , z0 ) on C. For this, we use the following result.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 77 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 78 / 98

3.8 Tangent Planes to Level Surfaces and Related Concepts 3.8 Tangent Planes to Level Surfaces and Related Concepts
Example 31
Find parametric equations of the tangent line to the curve of intersection
of the surfaces 3x2 + 2y 2 + z 2 = 49 and x2 + y 2 − 2z 2 = 10 at the point
Theorem (3, −3, 2).
Let C be a curve which is the intersection of two surfaces having equations

F (x, y, z) = 0 and G(x, y, z) = 0.

⇀ ⇀
Let P0 (x0 , y0 , z0 ) be a point on C, and N 1 and N 2 be normal vectors to the
surfaces F (x, y, z) = 0 and G(x, y, z) = 0, respectively, at the point P0 . If
⇀ ⇀
N 1 and N 2 are not parallel, then the direction numbers of the tangent line
⇀ ⇀
to the curve C at the point P0 are the components of the vector N 1 × N 2 .

Figure: Illustration for Example 31

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 79 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 80 / 98
3.8 Tangent Planes to Level Surfaces and Related Concepts 3.8 Tangent Planes to Level Surfaces and Related Concepts
If the normal vectors of the surfaces F (x, y, z) = 0 and G(x, y, z) = 0 at
Exercises 32
the point P0 are parallel (this occurs when the cross product of their normal
⇀ 1 Find an equation of the tangent plane and a set of parametric
vectors is 0 ), then it means that they have the same tangent plane at P0 .
equations of the normal line to the given surface at the indicated
In this case, we say that the two surfaces are tangent at P0 . point.
(a) (z 2 − x2 )xyz − y 5 = 5; (1, 1, 2)
(b) zx2 − xy 2 − yz 2 = 18; (0, −2, 3)
2 Find symmetric equations for the line tangent to the curve of
intersection of the surfaces

xyz = 1 and x2 + y 2 − z = 1

at the point (1, 1, 1).

3 Show that the symmetric equations of the normal line to the surface
x2 + 2y 2 − 3z 2 = 3 at the point (2, 1, −1) is
x−2 y−1 z+1
= = .
Figure: Tangent surfaces at P0 2 2 3
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 81 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 82 / 98

3.8 Tangent Planes to Level Surfaces and Related Concepts 3.9 Relative Extrema and the Second Derivative Test

Exercises 32 Definition
4 Show that the equation of the tangent plane to the elliptic paraboloid Let f be a function of two variables x and y. Then
x2 y 2 z 1 f has a relative maximum at the point (x0 , y0 ) if there exists an
+ 2 = open disk D centered at (x0 , y0 ) such that f (x0 , y0 ) ≥ f (x, y) for all
a2 b c
(x, y) in D;
at the point P0 (x0 , y0 , z0 ) is given by
2 f has a relative minimum at the point (x0 , y0 ) if there exists an
2x0 2y0 z + z0 open disk D centered at (x0 , y0 ) such that f (x0 , y0 ) ≤ f (x, y) for all
x+ 2 y = . (x, y) in D; and
a2 b c
3 f has a relative extremum at (x0 , y0 ) if f has either a relative
5 Prove that the surfaces 4x2 + y 2 + 9z 2 = 108 and xyz = 36 are
maximum or a relative minimum at (x0 , y0 ).
tangent at the point (3, 6, 2).

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 83 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 84 / 98
3.9 Relative Extrema and the Second Derivative Test 3.9 Relative Extrema and the Second Derivative Test

Definition
If f (x, y) exists at all points in some open disk B((x0 , y0 ); r), the point
(x0 , y0 ) is a critical point of f if one of the following conditions holds:
1 Both fx (x0 , y0 ) = 0 and fy (x0 , y0 ) = 0;
2 fx (x0 , y0 ) or fy (x0 , y0 ) does not exist.

Example 33
Find all critical points of the function f (x, y) = 6x − 4y − x2 − 2y 2 .

Figure: Relative extremum points of a function Definition

The point (x0 , y0 , f (x0 , y0 )) is a saddle point of z = f (x, y) if
Theorem 1 (x0 , y0 ) is a critical point of f ; and
Suppose that f (x, y) exists at all points in some open disk B((x0 , y0 ); r) 2 every open disk centered at (x0 , y0 ) contains points (x, y) in the
and f has a relative extremum at (x0 , y0 ). If fx (x0 , y0 ) and fy (x0 , y0 ) exist, domain of f in which f (x, y) < f (x0 , y0 ) and points (x, y) in the
then fx (x0 , y0 ) = 0 and fy (x0 , y0 ) = 0. domain of f for which f (x, y) > f (x0 , y0 ).
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 85 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 86 / 98

3.9 Relative Extrema and the Second Derivative Test 3.9 Relative Extrema and the Second Derivative Test

Theorem (Second-Derivative Test (SDT))

Let f be a function of two variables such that f and its first and second-
order partial derivatives are continuous on some open disk B((x0 , y0 ); r).
Suppose further that fx (x0 , y0 ) = 0 and fy (x0 , y0 ) = 0.
Let D = fxx (x0 , y0 )fyy (x0 , y0 ) − [fxy (x0 , y0 )]2 .
1 f has a relative minimum value at (x0 , y0 ) if

D>0 and fxx (x0 , y0 ) > 0 (or fyy (x0 , y0 ) > 0).

2 f has a relative maximum value at (x0 , y0 ) if

D>0 and fxx (x0 , y0 ) < 0 (or fyy (x0 , y0 ) < 0).
Figure: Saddle point of a graph of a function
3 f does not have a relative extremum, but (x0 , y0 , f (x0 , y0 )) is a
saddle point of f if D < 0.
4 No conclusion regarding relative extrema can be made if D = 0.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 87 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 88 / 98
3.9 Relative Extrema and the Second Derivative Test 3.9 Relative Extrema and the Second Derivative Test
Exercises 36
Example 34 1 Use the Second Derivative Test to determine the relative extrema and the
Use the Second Derivative Test to determine the relative extrema and the saddle points of the following functions, if there’s any.
saddle points of the following functions, if there’s any. (a) f (x, y) = 8 − 6y 2 − 6x2 + 6xy 2 + 2x3
1
2
f (x, y) = 6x − 4y − x − 2y 2 (b) g(x, y) = x3 + 3xy 2 − 3x2 − 3y2 + 4
1
2 g(x, y) = ex − x + ey − y (c) h(x, y) = x2 − x2 y + 2y 2
2
We apply SDT to some problems involving maximum or minimum values 2 Find the relative extrema and the saddle points of the function
of quantities.
f (x, y) = x2 + 1 − 2x cos y,
Example 35
where −π < y < π.
1 Find three positive numbers such that their product is 27 and their
3 A closed rectangular box with a volume of 16 ft2 is to be made of three
sum is as small as possible.
kinds of materials. The cost of the material for the top and the bottom is
2 Find the point on the surface z = x2 − y 2 + 2 that is closest to the Php 18.00 per square foot, the cost for the material for the front and back is
origin, and find the minimum distance. Php 16.00 per square foot, and the cost of the material for the other two
sides is Php 12.00 per square foot. Find the dimensions of the box such that
the cost of the materials is a minimum.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 89 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 90 / 98

3.10 Absolute Extrema and the Method of Lagrange Multipliers 3.10 Absolute Extrema and the Method of Lagrange Multipliers

Extrema problems have two types: Example 37

1 Free extrema problems: determining the extreme values of a Now, suppose that we will find the maximum value of f , but we impose
function in its domain that the variables x and y satisfy the following condition:
2 Constrained extrema problems: determining the extreme values of x + y = 2.
a function subject to certain condition/s, called constraints or side
conditions. We are now looking for a constrained maximum. The solution can be
found by subtituting y = 2 − x to the function f , to obtain a function F
Example 37
of one variable only:
Consider finding the maximum value of the function p
p F (x) = 25 − x2 − (2 − x)2 .
f (x, y) = 25 − x2 − y 2 .
One can use the methods of finding a maximum of functions of one
This is a free maximum problem, and using the Second Derivative Test, variable only, so that we obtain x = 1 and y = 1, and the constrained
one can verify that the maximum value of f occurs when (x, y) = (0, 0), maximum value is
with maximum value f (0, 0) = 5. p √
f (1, 1) = 25 − (1)2 − (1)2 = 23.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 91 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 92 / 98
3.10 Absolute Extrema and the Method of Lagrange Multipliers 3.10 Absolute Extrema and the Method of Lagrange Multipliers

Other than the Second Derivative Test, there is an alternative way to Remarks
find relative extrema of a function f of two variables x and y subject to the The method of Lagrange multipliers can be extended to a function f of
constraint g(x, y) = 0. This procedure is called the method of Lagrange three variables x, y and z subject to the constraint g(x, y, z) = 0. We only
multipliers. need to form the auxiliary function F (x, y, z, λ) = f (x, y, z) + λg(x, y, z),
Given a function f of x and y and a constraint g(x, y) = 0, we introduce a and solve the system 
new variable λ, called a Lagrange multiplier and form the auxiliary function Fx (x, y, z, λ) = 0


 F (x, y, z, λ) = 0
y
F (x, y, λ) = f (x, y) + λg(x, y).
 Fz (x, y, z, λ) = 0

Find the critical points of F , that is, solve the system


Fλ (x, y, z, λ) = 0

Fx (x, y, λ) = 0

Example 38
Fy (x, y, λ) = 0
 1 Use Lagrange multipliers to find the point on the surface
Fλ (x, y, λ) = 0

z = x2 − y 2 + 2 that is closest to the origin, and find the minimum
distance.
for the values of x, y and λ. The values of x and y that give the relative
extrema of f are among these critical points. 2 A box is to be constructed out of 96 square feet of material. Find the
dimensions that maximize the volume of the box.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 93 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 94 / 98

3.10 Absolute Extrema and the Method of Lagrange Multipliers 3.10 Absolute Extrema and the Method of Lagrange Multipliers

If the function f of variables x, y and z are subject to two constraints Exercises 39

g(x, y, z) = 0 and h(x, y, z) = 0, in addition to the variable λ, we introduce 1 A rectangular box, whose edges are parallel to the coordinate axes, is
another variable µ and form the auxiliary equation inscribed in the ellipsoid 36x2 + 4y 2 + 9z 2 = 36. What is the greatest
possible volume of the box?
F (x, y, z, λ, µ) = f (x, y, z) + λg(x, y, z) + µh(x, y, z).
2 Use Lagrange multipliers to find the point on the plane
Then we solve for the system 3x − 2y + z = 1 closest to the point (1, 0, −1).
 3 In a firm, the total production cost is 2x + 4y, where x units is the

 Fx (x, y, z, λ, µ) = 0 capital and y units is the labor cost. The production function is given

 Fy (x, y, z, λ, µ) = 0 by P (x, y) = 6xy + 2y 2 . Using Lagrange Multipliers, find the



Fz (x, y, z, λ, µ) = 0 maximum possible output P (x, y), if the total cost is 200 units?



 Fλ (x, y, z, λ, µ) = 0 4 Use Lagrange multiplier to find the dimensions and area of the largest



Fµ (x, y, z, λ, µ) = 0 rectangular lot that can be enclosed with a fixed length of fence M
meters long.
to find the critical values of F . The values of x, y and z that give the 5 The plane x + y + z = 12 intersects the paraboloid z = x2 + y 2 in an
relative extrema of f are among these values. ellipse. Find the point on the ellipse that is closest to the origin.
Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 95 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 96 / 98
Appendix: Derivatives of Some Functions Appendix: Derivatives of Some Functions

Derivatives of Some Functions Derivatives of Some Functions

(Here, u and v are differentiable functions of x, and a, c and r are constants with a > 0.)
(Here, u and v are differentiable functions of x, and a, c and r are constants with a > 0.)
Dx [c] = 0 Dx [uv] = u · Dx [v] + v · Dx [u] 1 −1
Dx (sin−1 u) = √ · Dx (u) Dx (cot−1 u) = · Dx (u)
Dx [cu] = c · Dx [u] hui v · Dx [u] − u · Dx [v] 1 − u2 1 + u2
Dx =
Dx [u ± v] = Dx [u] ± Dx [v] v v2 −1 1
Dx (cos−1 u) = √ · Dx (u) Dx (sec−1 u) = √ · Dx (u)
1 − u2 u u2 − 1
Dx (sin u) = cos u · Dx (u) Dx (cot u) = − csc2 u · Dx (u)
1 −1
Dx (cos u) = − sin u · Dx (u) Dx (sec u) = sec u tan u · Dx (u) Dx (tan−1 u) = · Dx (u) Dx (csc−1 u) = √ · Dx (u)
1 + u2 u u2 − 1
Dx (tan u) = sec2 u · Dx (u) Dx (csc u) = − csc u cot u · Dx (u)
Dx (ur ) = rur−1 · Dx (u) Dx (au ) = au ln a · Dx (u) Dx (sinh u) = cosh u · Dx (u) Dx (coth u) = − csch2 u · Dx (u)
1 1 Dx (cosh u) = sinh u · Dx (u) Dx (sech u) = − sech u tanh u · Dx (u)
Dx (ln u) = · Dx (u) Dx (loga u) = · Dx (u)
u u ln a
Dx (eu ) = eu · Dx (u) Dx (tanh u) = sech2 u · Dx (u) Dx (csch u) = − csch u coth u · Dx (u)

Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 97 / 98 Karlo S. Orge (MSU-IIT) MAT070 Chapter 3 98 / 98