(11) Partial Derivatives

11.1) Definition 1: PDs for Functions of Two Independent Variables

1

For a function z = f (x, y) of two independent variables x, y , the partial

derivatives fx and fy are functions defined by:
f ( x + h, y ) − f ( x , y )
f x ( x , y ) = lim
h→0 h
f ( x , y + h) − f ( x , y )
f y ( x , y ) = lim
h→0 h
Baby Notation for PDs: Given the function z = f (x, y), we may write:
f
f x ( x, y) = f x = z x = =  f ( x , y ) = z = f 1 = D x f = D1 f
x x x
f
f y ( x, y ) = f y = z y = =  f ( x , y ) = z = f 2 = D y f = D2 f
y y y
Example 1: In each part, compute the requested partial derivatives
a) For f (x, y) = x 3 + x 4y 5 – 5 y 2 compute fx (– 3, 2) and fy (2, – 1)
b) For f (x, y) = tan (16 – x – 4 y ) + ln (x 2 + y 2 ) compute fx and fy
x − y2
1 3
c) For f (x, y) = arcsin (x 2 y 3) + e 2
compute fx and fy
2 x5 y7
d) For f (x, y) = x 5 + y7 compute fx and fy

11.2) Interpretation of Partial Derivatives at Given Point P(a, b, f (a, b))

Figure 1 shows the graph of a 3D Figure 1 :Tangent Lines
surface defined by z = f (x, y)
1) The partial derivative z x = fx (a, b)
is the slope of tangent line L1 to
curve C1 : z = f (x, b) at the point
P(a, b, f (a, b))
2) The partial derivative z y = fy (a, b)
is the slope of tangent line L2 to
curve C2 : z = f (a, y) at the point
P(a, b, f (a, b))

(a, b, 0)
Figure 1A: Tangent Lines to a 3D Surface
z

x
Example 2: For the function f (x, y) = 6 – x 2 – y 2
a) Evaluate and interpret f x 32 ,1 ( ) ( )
b) Evaluate and interpret f y 32 ,1
c) Write the parametric form of the equation of the line tangent to the
curve f (x, 1), which is the curve of intersection of the given surface with
the plane y = 1
d) Write the parametric form of the equation of the line tangent to the
( )
curve f 32 , y , which is the curve of intersection of the given surface
with the plane x = 32
e) Sketch the surface and the tangent lines.
Answers: a) f x ( x , y ) = −2 x ( )
f x 32 ,1 = −3 Interpret
b) f y ( x , y ) = −2 y fy ( 32 ,1) = −2 Interpret
c) f (x, 1) = 5 – x 2 Space Curve: C1 = t, 1, 5 – t 2
(
Point of Tangency: P 32 ,1, 11
4 )
Realize: t = 32
Direction Vector: C1 = 1, 0, –2t  = 1, 0, –3 
Tangent Line L1: x = 32 + t ; y = 1 + 0t ; z = 11
4 – 3t
( )
d) f 32 , y = 15
4 –y
2
4 −t
Space Curve: C2 = 32 , t , 15 2

(
Point of Tangency: P 32 ,1, 114 )Realize: t = 1
Direction Vector: C 2 = 0, 1, –2t  = 0, 1, –2 
Tangent Line L2: x = 32 + 0t ; y = 1 + t ; z = 11
4 – 2t
Example 3: For the function f (x, y) = 4x 2y 3, compute f x ( x, y ) and f y ( x , y )
by using the definitions.
f ( x + h, y ) − f ( x , y )
f x ( x , y ) = lim
h→0 h
f ( x , y + h) − f ( x , y )
f y ( x , y ) = lim
h→0 h
Answer: f x ( x, y ) = 8x y 3 ; f y ( x , y ) = 12x 2y 2
Example 4: In each part, find the first partial derivatives for the given functions
y x
a) f (x, y) = x 9y 10 – 10 x 2y – cot( x y) b) f ( x , y ) =
(2 x−5 y )2
c) f ( x , y ) = tan −1 ( xy 2 ) + e
− 1( x 2 + y3 )
2
( y)
d) f ( x , y , z ) = 4 x 3 y sec −1 z
e) f ( x, y, z ) = cos(2 x 2 y 3 z 4 ) + 3 x −1 + 3 y − z 2

11.3) Second Partial Derivatives

  f   2z  2 f
( z x ) x = ( f x ) x = z xx = f xx ( x, y ) = f11 = x  x  = x2 = x2
  f   2z 2 f
( z x ) y = ( f x ) y = z xy = f xy ( x, y ) = f12 = y  x  = yx = yx
f   2z 2 f
( ) ( )  
z y = f y = z yx = f yx ( x , y ) = f 21 = x  y  = xy = xy
x x  
f   2z  2 f
( ) ( ) y y
 
z y = f y = z yy = f yy ( x, y ) = f 22 = y  y  = 2 = 2
  y y
Example 5: In each part, find all second partial derivatives for the given functions
y x
a) f (x, y) = x 9y 10 – 10 x 2y – cot( x y) b) f ( x , y ) =
(2 x−5 y )2
− 1( x 2 + y3 )
c) f ( x , y ) = tan −1 ( xy 2 ) + e 2 d) ln(x 2 + y 2)

11.4) Clairut’s Theorem 1: Equality of Second Mixed Partial Derivatives

Suppose function f is defined on an open set D and f xy , f yx are both
continuous at every point in D. Then f xy = f yx at every point in D.
Example 6: Verify Clairut’s Theorem in each part of Example 5 .
11.5) Partial Derivatives of Third and Higher Order
Partial Derivatives of order three or higher can be defined as well.
 2 f  3 f
( )
For instance: f xyx = f xy = 
x
 =
x  yx  xyx

z xxy = ( z xx ) y =    2
z  =  3
z
 
y  x  yx 2
2

 3 f  4 f
(
f xyyy = f xyy = ) y
 =
y  y 2x  y 3x
 3 
( )
z xyyx = z xyy =   2 z  =  z2
4

x x  y x  xy x

Baby Remark: Using Clairut’s Theorem, we can show that

f xxy = f xyx = f yxx and f xyy = f yxy = f yyx
if these functions are continuous on an open set D.
Example 7: In each part, find the indicated partial derivatives for the functions
a) Given f (x, y) = x 5y 7 – x 4y 6 + ex y, verify that
f xxy = f xyx = f yxx and f xyy = f yxy = f yyx
b) Given z = cos (5y – x z 3), find: b1) f zyyx b2) f yxzyz

11.6) Partial Differential Equations

Partial Derivatives are often used to describe many physical laws
through partial differential equations.
Some partial differential equations have particular names.
Example 8: Examples of popular partial differential equations
a) Transport Equation: u + c u = 0 b) Laplace Equation: uxx + uy y = 0
t x
1

u   u  2u 
2
 2 u u
c) Heat Equation: =  2 + 2 
d) + = u( x , t )
t   x  y   x 2
 t
 2u 2  u  2u  u v w
2 1
e) Wave Equation: 2 =   2 + 2  f) + + = ( xyz ) 3
t  x y  x y z
h 2  2 ( x , t )  ( x , t )
g) Schrödinger Equation: U ( x )( x , t ) − = ih
2m x 2 t
Example 9: Show that the function u( x , y ) = e x sin y is a solution of Laplace
Equation. Solutions to Laplace Equation are called Harmonic Functions
and are of paramount importance in the analysis of heat conduction, fluid
flow dynamics, electric potential, gravitation, etc…
Example 10: Show that the function u(x, t) = sin(x –  t) is a solution of the
Wave Equation. This equation describes the motion and dynamics of
waveforms, such as ocean waves, sound waves, light waves, waves
traveling on a vibrating string (like guitar strings, violin strings).
Example 11: Cobb-Douglas Production Function P(L, K) = b L K 1 – 
provides a simplified description of economy performance in which
which production output is determined by the amount of labor input L
and the amount of capital input K invested.
a) Compute and interpret the marginal productivity of labor P
L
b) Compute and interpret the marginal productivity of capital P
K
11.7) Tangent Planes
Figure 3 shows the graph of a 3D Figure 3 :Tangent Plane
surface defined by z = f (x, y) and
the associated tangent plane at
point P(x0, y0, f (x0, y0)).
Equation of plane containing P :
A(x – x0) + B(y – y0) + C(z – z0) = 0
Divide by C  0 ; let a = –A/C and
let b = –B/C to obtain:
z – z0 = a(x – x0) + b(y – y0)
1) When y = y0 , curve C1 : z = f (x, y0)
has tangent line L1 whose equation
z – z0 = a(x – x0) shows slope
a = (z – z0)/(x – x0) = fx (x0, y0)
2) When x = x0 , curve C2 : z = f (x0, y) has tangent line L2 whose equation
z – z0 = b(y – y0) shows slope b = (z – z0)/(y – y0) = fy (x0, y0)
3) Equation of Tangent Plane:
z – z0 = fx (x0, y0) (x – x0) + fy (x0, y0) (y – y0)
Example 12: Verify the tangent plane for the function of Example 2
the function f (x, y) = 6 – x 2 – y 2 where the resulting tangent lines
were: L1: x = 32 + t ; y = 1 + 0t ; z = 11 4 – 3t and
L2: x = 32 + 0t ; y = 1 + t ; z = 11 3 11
4 – 2t at P 2 ,1, 4 ( )
Explain any discrepancies.
Example 13: In each part, find an equation of the tangent plane the given
surface at the specified point. Use the gradient vector.
a) z = 3y2 – 2x2 + x at P (2, – 1, – 3) Answer: z = –7x – 6y + 5
b) z = x sin(x + y) at P (– 1, 1, 0) Answer: z = –x – y
c) z = x2 + xy + 3y2 at P (1, 1, 5) Answer: z = 3x + 7y – 5
d) z = arctan(xy2) at P 1,1, 4 ( )
Answer: z = 12 x + y – 32 + 4

11.8) Definition 2: The Gradient Vector 1

a) For Functions z = f (x, y) of Two Independent Variables x, y ,

the gradient of f is the vector function f defined by
f f
f ( x, y ) = f x ( x, y ), f y ( x, y ) = f x ( x, y ) i + f y ( x, y ) j = i+ j
x y
b) For Functions w = f (x, y, z) of Three Independent Variables x, y, z ,
the gradient of f is the vector function f defined by
f ( x , y , z ) = f x ( x , y , z ), f y ( x , y , z ), f z ( x , y, z ) or
f f f
f ( x, y, z ) = f x ( x , y, z ) i + f y ( x , y, z ) j + f z ( x, y, z ) k = i+ j+ k
x y z
More Baby Remarks:
1) The symbol  is called “del operator”
2) The expression f is read “gradient of f ” or “grad f ” or “del f ”

Example 14: In each part, find the gradient vector of given function z = f (x, y)
a) z = 3y2 – 2x2 + x Answer: f ( x , y ) = ( −2 x + 1) i + 6 y j
b) f (x, y) = x sin(yz) – 2e xyz
Answer:
f ( x, y, z ) = sin( yz ) − 2 yz e xyz , xz cos( yz ) − 2 xz e xyz , xy cos( yz ) − 2 xy e xyz
11.9) Idea
The idea of Definition 2 derived from the fact that the equation of the
tangent plane to the surface z = f (x, y) is the dot product of two vectors.
Equation of Tangent Plane: z – z0 = fx (x0, y0) (x – x0) + fy (x0, y0) (y – y0)
Rewrite: fx (x0, y0) (x – x0) + fy (x0, y0) (y – y0) – (z – z0) = 0
Or: fx (x0, y0), fy (x0, y0), –1 • (x – x0), (y – y0), (z – z0) = 0
Notice that equation f (x, y, z) = c (a constant c   ) represents a level
surface S of a function of three variables w = f (x, y, z). If P(x0, y0, z0) is a
point on S and C is a curve that lies on S and passes through P, we can
describe curve C by a continuous vector function r (t) = x (t), y (t), z (t)
with P(x0, y0, z0) being: r (t0) = x (t0), y (t0), z (t0). Because curve C lies
on surface S, then any point (x (t), y (t), z (t)) on C satisfies the equation
of S. So: f (x (t), y (t), z (t)) = c (a constant c   ).
If f is a differentiable function and x, y, z are all three differentiable
functions of t, then we apply the Chain Rule of differentiation to both
sides of the equation f (x (t), y (t), z (t)) = c
f dx f dy f dz
Obtain: + + =0
x dt y dt z dt
f f f dy dz
Since f = , , and r ( t ) = dx , , , we recognize
x y z dt dt dt
the dot product f • r ( t ) = 0 and when t = t0 , we get:
f ( x0 , y0 , z 0 ) • r ( t 0 ) = 0
This means that gradient vector f ( x0 , y0 , z 0 ) at P(x0, y0, z0) is
perpendicular to the tangent vector r ( t 0 ) to any curve C on surface S
that passes through any point P(x0, y0, z0).
So, the tangent plane to the level surface f (x, y, z) = c is:
fx (x0, y0, z0) (x – x0) + fy (x0, y0, z0) (y – y0) + fz (x0, y0, z0) (z – z0) = 0,
where fz (x0, y0, z0) (z – z0) = –1 by rewriting z = f (x, y) as
f (x , y ) – z = 0
11.10) Definition 3: Directional Derivative of a Function z = f (x, y) in the
Direction of a Unit Vector u = a, b
The directional derivative of function f at (x0, y0) in the direction of a
unit vector u = a, b is defined by
𝒇(𝒙𝟎 + 𝒉𝒂, 𝒚𝟎 + 𝒉𝒃) − 𝒇(𝒙𝟎 , 𝒚𝟎 )
Du f (x0, y0) = 𝒍𝒊𝒎
𝒉→𝟎 𝒉
if the limit exists and is finite.
11.11) Theorem 2: Direct Computation of a Directional Derivative
If f is a differentiable function in the independent variables x, y, then
f has a directional derivative in the direction of any unit vector
u = a, b and Du f (x, y) = f x (x, y) a + f y (x, y) b
Realize that: Du f (x, y) = f (x, y) • u
Corollary 1: More Computation of a Directional Derivative
If the unit vector u makes an angle  with the positive x-axis, then we
write u = cos , sin   and the directional derivative from Theorem 2
becomes: Du f (x, y) = fx (x, y) cos  + fy (x, y) sin 
Example 15: Find the directional derivative of f (x, y) = 5x 2y 3 – 4x + 2 y at
(–2, 1) in the direction of vector v = – i + 4 j
Solution: Invoke and apply Theorem 2 Du f (x, y) = f (x, y) • u
Where: f (x, y) = (10x y 3 – 4) i + (15 x 2 y 2 + 2) j = – 24 i + 62 j
𝒗 𝟏 𝟒
u= =– i+ j
|𝒗| √𝟏𝟕 √𝟏𝟕
𝟏 𝟒
Therefore: Du f (–2, 1) = (– 24 i + 62 j ) • (– i+ j )
√𝟏𝟕 √𝟏𝟕

𝟐𝟕𝟐
Obtain: Du f (–2, 1) =
√𝟏𝟕
 𝟐𝝅
Example 16: If u is the unit vector given by angle  =
𝟑
a) Find the directional derivative Du f (x, y) for f (x, y) = 4y 3 – 3x 2 + 5x y
b) Find Du f (2, –1) for the same f (x, y) = 4y 3 – 3x 2 + 5x y
Solution: a) Invoke and apply Corollary 1
𝟐𝝅 𝟐𝝅
Du f (x, y) = f x (x, y) cos + fy (x, y) sin
𝟑 𝟑
−𝟏 √𝟑
Du f (x, y) = (– 6x + 5 y)· + (12y 2 + 5 x)·
𝟐 𝟐
2 𝟓 𝟔+𝟓√𝟑
Obtain: Du f (x, y) = 𝟔√𝟑 y – 𝟐y+ 𝟐
x
Solution: b) Invoke and apply the result from part a)
𝟏𝟕+𝟐𝟐√𝟑
Obtain: Du f (2, –1) =
𝟐
11.12) Definition 4: Directional Derivative of Functions of Three Variables
w = f (x, y, z) in the Direction of Unit Vector u = a, b, c
The directional derivative of function f at (x0, y0, z0) in the direction of
a unit vector u = a, b, c is defined by
𝒇(𝒙𝟎 + 𝒉𝒂, 𝒚𝟎 + 𝒉𝒃, 𝒛𝟎 + 𝒉𝒄) − 𝒇(𝒙𝟎 , 𝒚𝟎 , 𝒛𝟎 )
Du f (x0, y0, z0) = 𝒍𝒊𝒎
𝒉→𝟎 𝒉
if the limit exists and is finite.
11.13) Theorem 3: Direct Computation of a Directional Derivative
If f is a differentiable function in the independent variables x, y, z, then
f has a directional derivative in the direction of any unit vector
u = a, b, c and Du f (x, y, z) = f x (x, y, z) a + f y (x, y, z) b + f z (x, y, z) c
Realize that: Du f (x, y, z) = f (x, y, z) • u
xy
Example 17: Find the directional derivative of f (x, y, z) = 3z e at
(–2, 1, –1) in the direction of vector v = – i + 2 j + 3 k
𝟑𝒆+𝟑𝒆−𝟐
Answer: Du f (–2, 1, –1) = f (–2, 1, –1) • u =
√𝟏𝟒
11.14) Theorem 4: Maximization of the Directional Derivative
If f is a differentiable function of two or three or more independent
variables X = x1 , x2 , x3, …, xn , then the maximum value of the
directional derivative Du f (X) is |  f (X) | and occurs when unit vector
u has the same direction as the gradient vector  f (X). So, maximum
rate of change of f is the magnitude of the gradient vector  f (X)
1

when unit vector u has the same direction as gradient vector  f (X).
Proof: We learned that Du f = f • u = |  f (X) | | u | cos  where
 is the angle between f and u
The maximum value of cos  is and it occurs when  = 0
Since | u | = 1, then the maximum value of Du f is |  f (X) | and it
occurs when  = 0, which is when u has the same direction as f
xy
Example 18: a) Find the rate of change of f (x, y) = 4x2 y2 + 2 e at P(–2, 1)
in the direction from P to Q(1, 3)
b) Find the maximum rate of change of f
c) In what direction does f has maximum rate of change?
Solution: a) First, compute the gradient vector
f (x, y) =  fx , fy =  8x y2 + 2 y e x y , 8x2 y + 2 x e x y 
f (–2, 1) =  –16 + 2 e –2 , 32 – 4 e –2 
Next, compute unit vector u in the direction of v = PQ = 3 , 2
u= v = 3 , 2
|v | 13 13
Next, compute the rate of change of f in the direction from
1

−2
P to Q Du f (–2, 1) = f (–2, 1) • u = 16 − 2e
13
b) Maximum rate of change of f is
|  f (–2, 1) | = 1280 − 320e −2 + 20e −4
c) In the direction of gradient vector
f (–2, 1) =  –16 + 2 e –2 , 32 – 4 e –2 