Partial Derivatives

Applied Mathematics MAT538

1 Introduction
In this section, we will learn how to differentiate a function of several independent variables. Let f (x, y) be the
function of two independent variables, then the first-order partial derivatives or first partial derivatives
with respect to x and y at point (x0 , y0 ) can be denoted as

∂f ∂f
(x0 , y0 ) = fx (x0 , y0 ), and (x0 , y0 ) = fy (x0 , y0 ).
∂x ∂y
In partial derivative, when we differentiate the function with respect to any of independent variable, other
independent variable(s) will be treated as constant(s).
∂f ∂f
Example 1. Find the value of and at the point (4, −5) if f (x, y) = x2 + 3xy + y − 1.
∂x ∂y
Solution

∂f ∂ ∂f ∂
x2 + 3xy + y − 1 ;


x2 + 3xy + y − 1 ;


= =
∂x ∂x ∂y ∂y
= 2x + 3(1)(y) + 0 − 0; = 0 + 3(x)(1) + 1 − 0;
= 2x + 3y. = 3x + 1.
∂f ∂f
Thus (4, −5) = fx (4, −5) = 2(4) + 3(−5) = −7. Thus (4, −5) = fy (4, −5) = 3(4) + 1 = 13.
∂x ∂y
y
Example 2. Find fy if f (x, y) = y sin(xy) + (x + y)6 + .
y+1
Solution
 
∂ ∂ ∂ y
fy = [y sin(xy)] + (x + y)6 + ;
∂y ∂y ∂y y+1
∂ ∂
(y + 1) (y) − (y) (y + 1)
∂ ∂ ∂ ∂y ∂y
= y [sin(xy)] + sin(xy) [y] + 6(x + y)5 (x + y) + ;
∂y ∂y ∂y (y + 1)2
| {z } | {z } | {z }
product rule chain rule quotient rule
 
 (y + 1)(1) − (y)(1 + 0)
5

= y cos(xy)(x) + sin(xy)(1) + 6(x + y) (0 + 1) + ;
(y + 1)2
1
= xy cos(xy) + sin(xy) + 6(x + y)5 + .
(y + 1)2

Practice
Find all the first partial
 derivatives for the following functions
1. f (x, y) = 2x2 − 3y − 4. ans
fx = 4x, fy = −3
 1 x
2. f (x, y) = x/y. ans
fx = , fy = − 2 .
y y
p  x y
3. f (x, y) = x2 + y 2 . ans
fx = p , fy = p .
2
x +y 2 x + y2
2
 exy
4. f (x, y) = exy ln(y). ans
fx = yexy ln(y), fy = + xexy ln(y).
y
2y − x 
5. f (x, y) = , (−1, 1). ans
fx = −4, fy = −1.
x+y+1

Page 1 of 7
Partial Derivatives
Applied Mathematics MAT538

1.1 Implicit Differentiation

∂z ∂z
Example 3. Let z = f (x, y), find and for the equation yz 2 − ln(z) = x + y.
∂x ∂y
Solution

∂  2 ∂ ∂ ∂ ∂  2 ∂ ∂ ∂
yz − [ln(z)] = [x] + [y]; yz − [ln(z)] = [x] + [y];
∂x ∂x ∂x ∂x ∂y ∂y ∂y ∂y
∂z 1 ∂z ∂z 1 ∂z
y(2z) − = 1 + 0; y(2z) + z 2 (1) − = 0 + 1;
 ∂x z ∂x ∂y z ∂y
1 ∂z
 
1 ∂z
2yz − = 1; 2yz − = 1 − z2;
z ∂x z ∂y
∂z 1 1 − z2 ∂z
= . . =
∂x 1 1 ∂y
2yz − 2yz −
z z
Alternatively, we can use Implicit Function Theorem to do the differentiation. The theorem state that if
z = f (x, y) is differentiable and F (x, y, z) = 0 then

∂z ∂F ∂F Fx ∂z ∂F ∂F Fy
zx = =− / =− , zy = =− / =− .
∂x ∂x ∂z Fz ∂y ∂y ∂z Fz
∂z ∂z
Example 4. Use Implicit Function Theorem to find and for the equation in example 3.
∂x ∂y
Solution
Let F (x, y, z) = yz 2 − ln(z) − x − y.
Finding ∂z/∂x then Finding ∂z/∂y then,
1 1
Fx = −1, Fz = 2yz − Fy = z 2 − 1, Fz = 2yz −
z z
therefore therefore
   
∂z Fx −1 1 ∂z Fy z2 −1  1 − z2
=− = − = . =− = − = .
  
∂x Fz 1 1 ∂y Fz 1  1
2yz − 2yz − 2yz − 2yz −
z z z z
∂z
Example 5. If e1+z + xy 2 + z = 1, find . (Jan 2018-4 marks)
∂x
Solution
By ordinary implicit differentiation, By Implicit Function Theorem,
Let F (x, y, z) = e1+z + xy 2 + z − 1 then
∂  1+z ∂
+ xy 2 + z = 1

e = [1];
∂x  ∂x Fx = y 2 , Fz = e1+z + 1.
1+z ∂z ∂z
e 0+ + y2 + = 0;
∂x ∂x therefore
∂z ∂z 
y2

−y 2
e1+z + = −y 2 ; ∂z
=−
Fx
= − 1+z = 1+z .
∂x
 1+z ∂x ∂x Fz e +1 e +1
= −y 2 ;

e +1
∂z −y 2
= .
∂x e1+z + 1

Page 2 of 7
Partial Derivatives
Applied Mathematics MAT538

Practice
Let z = f (x, y). Use ordinary implicit differentiation and Implicit Function Theorem to find ∂z/∂x
and ∂z/∂y for the following implicit equations
 x −2y
1. x2 + 2y 2 + 3z 2 = 1. ans
zx = − , zy = .
 yz 3z xz 3z
2. ez = xyz. ans
zx = x , zy = z .
e − xy  e − xy


1 + ln(x) −z
3. yz + x ln x = z 2 . ans
zx = − , zy = .
y − 2z y − 2z
 e cos(x) − 6(y + z)(xz + xy)5
x
 y
e sin(x) − 6x(xz + xy)5
 
y 6
4. e sin(x) = (xz + xy) . ans
zx = , zy = .
6x(xz + xy)5 6x(xz + xy)5

1.2 Higher Partial Derivatives

Let z = f (x, y).
Second Partial Derivatives
∂2z ∂2f
 
∂ ∂f
= (x, y) = (x, y) = [fx (x, y)]x = fxx (x, y).
∂x2 ∂x2 ∂x ∂x

∂2z ∂2f
 
∂ ∂f
= (x, y) = (x, y) = [fy (x, y)]y = fyy (x, y).
∂y 2 ∂y 2 ∂y ∂y
Mixed Partial Derivatives
∂2z ∂2f
 
∂ ∂f
= (x, y) = (x, y) = [fy (x, y)]x = fyx (x, y).
∂x∂y ∂x∂y ∂x ∂y

∂2z ∂2f
 
∂ ∂f
= (x, y) = (x, y) = [fx (x, y)]y = fxy (x, y).
∂y∂x ∂y∂x ∂y ∂x
Example 6. Find fxx , fyy , fxy and fyx for f (x, y) = x3 + x2 y 3 − 2y 2 .
Solution    
∂ ∂f ∂ ∂ ∂f ∂
fxx = = [3x2 + 2xy 3 ] = 6x + 2y 3 . fxy = = [3x2 + 2xy 3 ] = 6xy 2 .
∂x  ∂x ∂x ∂y  ∂x  ∂y
∂ ∂f ∂ ∂ ∂f ∂
fyy = = [3x2 y 2 − 4y] = 6x2 y − 4 . fyx = = [3x2 y 2 − 4y] = 6xy 2 .
∂y ∂y ∂y ∂x ∂y ∂x

Example 7. If f (x, y) = ex cos(y), verify that fxy = fyx . (Jun 2018-3 marks)
Solution
 
∂ ∂f ∂ x
fxy = = [e cos(y)] = −ex sin(y).
∂y ∂x ∂y
 
∂ ∂f ∂
fyx = = [−ex sin(y)] = −ex sin(y).
∂x ∂y ∂x
Since fxy = fyx , then the equation is verified.
Practice
Find all the second partial derivatives.
1. f (x, y) = x + y + xy.

ans
fx = 1 + y, fxx = 0, fy = 1 + x, fyy = 0, fxy = 1, fyx = 1.
2. g(x, y) = x2 y + cos(y) + y sin(x).
 2
ans
gx = 2xy + y cos(x), gxx = 2y − y sin(x), gy = x − sin(y) + sin(x), gyy = − cos(y), gxy = 2x + cos(x),
3. r(x, y) = ln(x + y).
 1 1 1 1 1 1
ans
rx = x + y , rxx = − (x + y)2 , ry = x + y , ryy = − (x + y)2 , rxy = − (x + y)2 , ryx = − (x + y)2 .
4. w(x, y) = x sin(xy).
 2 2
ans
wx = xy cos(xy) + sin(xy), wxx = −xy sin(xy) + 2y cos(xy), wy = x cos(xy),
3 2 2
wyy = −x sin(xy), wxy = −x y sin(xy) + 2x cos(xy), wyx = −x y sin(xy) + 2x cos(xy)..

Page 3 of 7
Partial Derivatives
Applied Mathematics MAT538

2 Chain Rule
Suppose that z = f (x, y) is a differentiable function of x and y where x = g(t) and y = h(t) are both
differentiable functions of t, then z is a differentiable function of t,
dz ∂f dx ∂f dy
= + .
dt ∂x dt ∂y dt
dz
Example 8. If z = x2 y + 3xy 4 , where x = sin(2t) and y = cos(t), find when t = 0.
dt
Solution
The chain rule gives
dz ∂z dx ∂z dy
= + ;
dt ∂x dt ∂y dt
= (2xy + 3y 4 )[2 cos(2t)] + (x2 + 12xy 3 )[− sin(t)];
= 2 sin(2t) cos(t) + 3 cos4 (t) [2 cos(2t)] + sin2 (2t) + 12 sin(2t) cos3 (t) [− sin(t)].
   

when t = 0,
dz
= (0 + 3)[2 cos(0)] + (0 + 0)[− sin(0)] = (3)(2) = 6.
dt
Suppose that z = f (x, y) is a differentiable function of x and y, where x = g(s, t) and y = h(s, t) are
differentiable functions of s and t, then
∂z ∂z ∂x ∂z ∂y ∂z ∂z ∂x ∂z ∂y
= + , = + .
∂s ∂x ∂s ∂y ∂s ∂t ∂x ∂t ∂y ∂t
∂z
Example 9. If z = ex ln(y), where x = st2 and y = s2 t. Find and ∂z/∂t.
∂s
Solution
The chain rule gives

∂z ∂z ∂x ∂z ∂y ∂z ∂z ∂x ∂z ∂y
= + , = + ;
∂s ∂x ∂s ∂y ∂s ∂t ∂x ∂t ∂y ∂t
    x
1 x e
= [ex ln(y)][t2 ] + e [2st]; x
= [e ln(y)][2ts] + [s2 ];
y y
2ex st ex s2
= ex t2 ln(y) + . = 2tsex ln(y) + .
y y
∂z 2
Example 10. If z = ln(x2 + y 2 ), x = r cos(t) and y = r sin(t). Show that = .
∂r r
(Jan 2018-5 marks)
Solution
∂z 2x ∂z 2y ∂x ∂y
= 2 ; = 2 ; = cos(t); = sin(t).
∂x x + y2 ∂y x + y2 ∂r ∂r

∂z ∂z ∂x ∂z ∂y
= + ;
∂r ∂x ∂r ∂y ∂r
2x cos(t) 2y sin(t)
= + 2 ;
x2 + y 2 x + y2
2r cos(t) cos(t) + 2r sin(t) sin(t)
= ;
r2 (cos2 (t) + sin2 (t))
2r[cos2 (t) + sin2 (t)] 2
= = .
r2 [cos2 (t) + sin2 (t)] r

Page 4 of 7
Partial Derivatives
Applied Mathematics MAT538

Practice dz ∂z ∂z
Use the Chain Rule to find , and .
dt ∂s ∂t
1. z(x, y) = xy 3 − x2 y, x = t2 + 1, y = t2 − 1
 dz 3 2 2
ans
dt = 2t(y − 2xy + 3xy √− x )
2. z(x, y) = sin(x) cos(y), x = t, y = 1t
 dz 1 1
ans
dt = √ cos(x) cos(y) + t2 sin(x) sin(y)
2 t
3. z(x, y) = (x − y)5 , x = s2 t, y = st2
 ∂z 4 2 ∂z 4 2
ans
∂s = 5(x − y) (2st − t ), ∂t = 5(x − y) (s − 2st).
4. z(x, y) = ln(3x + 2y), x = s sin(t), y = t cos(s)
 ∂z 3 sin(t) − 2t sin(s) ∂z 3s cos(t) + 2 cos(s)
ans
∂s = 3x + 2y
,
∂t
=
3x + 2y
.

3 Extreme Values and Saddle Point

In this section, we will learn how to find critical values and its type of multi-variable function.
Second Derivative Test for Local Extreme Values
Suppose that f (x, y) and its first and second partial derivatives are continuous throughout a disk centered at
(a, b) and that fx (a, b) = fy (a, b) = 0, and the discriminant or Hessian of f , D is

D = fxx (x, y)fyy (x, y) − [fxy (x, y)]2 .

then

(a) f has a local maximum at (a, b) if fxx < 0 and D > 0 at (a, b).

(b) f has a local minimum at (a, b) if fxx > 0 and D > 0 at (a, b).

(c) f has a saddle point at (a, b) if D < 0.

(d) The test in inconclusive if D = 0.

Example 11. Find the local extreme values of the function

f (x, y) = xy − x2 − y 2 − 2x − 2y + 4.

Solution
First, we list the related higher derivatives,

fx = y − 2x − 2, fxx = −2, fy = x − 2y − 2, fyy = −2, fxy = 1.

Finding the critical points, let fx = fy = 0, therefore

y − 2x − 2 = 0 and x − 2y − 2 = 0.

Critical point after solving above equations is (−2, −2). The point is the only point where f may take on an
extreme value. Now, we will determine the locality type of critical point by using the Second Derivatives Test.
For the purpose, we will construct a table as follows,

Critical point D fxx Type of critical point

(-2,-2) 3 (<0) -2 (>0) local maximum.
(−2)(−2) − [1]2 = −2

Therefore (-2,-2) is a local maximum for the function f .

Page 5 of 7
Partial Derivatives
Applied Mathematics MAT538

Example 12. Find all the critical points of the function f (x, y) = x3 + y 3 + 3x2 − 3y 2 − 8 and classify each
as relative maxima, relative minima and saddle point. (Jan 2018 : 11 marks)

Solution
Let
fx = 3x2 + 6x, fxx = 6x + 6, fy = 3y 2 − 6y, fyy = 6y − 6, fxy = 0
.
Finding the critical points, let fx = fy = 0 then 3x2 + 6x = 0 and 3y 2 − 6y = 0.

3x2 + 6x = 0; 3y 2 − 6y = 0;
3x(x + 2) = 0; 3y(y − 2) = 0;
x = 0, x = −2 y = 0, y = 2
Thus, the critical points are (0,0), (0.2), (-2,0) and (-2,2).
Determine the type of critical points by using Second Derivative Test

Critical point D fxx Type of critical point

(0,0) -36 (<0) - saddle.
(0,2) 36 (>0) 6 (>0) local minimum.
(-2,0) 36 (>0) -6 (<0) local maximum.
(-2,2) -36 (<0) - saddle.

Practice
Find the local maximum, minimum or saddle
point(s) of the functions.

1. f (x, y) = 9 − 2x + 4y − x2 − 4y 2 . ans
max −1, 12

2. f (x, y) = y 3 + 3x2 y − 6x2 − 6y 2 + 2.  ans
max(0,0), min
 (0,4), saddle (2,2), saddle (-2,2)


1 1 1 1
3. f (x, y) = x2 + y 4 + 2xy. ans
min √ , − √ , min − √ , √ , saddle (0,0).
 2 2 2 2
4. f (x, y) = x3 − 3x + 3xy 2 . ans
max (-1,0), min (1,0), saddle (0,1), saddle (0,-1).

Page 6 of 7
 Partial Derivatives
Applied Mathematics MAT538

Final Examination Questions

Jul2019 (a) Find and sketch the domain for f (x, y) = ln x2 − y . (3 marks)


∂z ∂z
(b) Find and if eyz − 2xz 2 + sin(xy) = 10y defines z implicitly as a function of x and y. (8 marks)
∂x ∂y
(c) Find all the critical points of the function f (x, y) = 8y 3 +x3 +6xy and classify each as relative extrema
or saddle point. (11 marks)
∂z 10 − zeyz − x cos(xy) ∂z 2z 2 − y cos(xy)
ans: (a) Df = {(x, y)|x2 − y > 0}. (b) = = . (c)
  ∂y yeyz − 4xz ∂x yeyz − 4xz
1
(0, 0)saddle, −1, − maxima.
2
p  ∂z ∂z
Dec2018 (a) If z = ln x2 + y 2 , show that x +y = 1. (5 marks)
∂x ∂y
∂z
(b) Given yx2 + cos(xy) + e3xz = 1, find . (5 marks)
∂y
(c) Find all critical points of the function f (x, y) = xy − x3 − 2y 2 . Classify them as a relative maxima, a
relative minima, or a saddle point. (10 marks)
 
∂z sin(xy) − x 1 1
ans: (b) = . (c) (0, 0)saddle, , minimum.
∂y 3e3xz 12 48
dz
Jul2017 (a) If z = sin(x + y), x = t−2 , y = t2 , compute using chain rule. (4 marks)
dt
∂z ∂z
(b) If x2 − y 2 + z 2 − 2z = 4, evaluate the partial derivatives and . (5 marks)
∂y ∂x
5 4
(c) Find all critical points of the function f (x, y) = x + y − 5x − 32y − 3 and classify each as relative
maxima, relative minima or saddle point. (10 marks)
∂z y ∂z x
ans: (a) (2t − 22t−3 ) cos(t−2 + t2 ). (b) = , =− . (c) (−1, 2)saddle, (1, 2)minimum.
∂y z−1 ∂x z−1
Dec2016 (a) If f (x, y) = ex cos(y), verify fxy = fyx . (3 marks)
∂z
(b) Evaluate at (-2,2,-1) for x2 y + xz + z 2 = 4. (5 marks)
∂x
(c) Find the critical points of the function f (x, y) = x3 − y 2 − xy + 1 and identify the nature of each
point. (11 marks)
 
∂z 1 1
ans: (b) = −9/4. (c) (0, 0)saddle, − , minimum.
∂x 6 12
Jun2016 (a) If f (x, y) = ex cos(y), verify fxy = fyx . (3 marks)
∂z
(b) Evaluate at (-2,2,-1) for x2 y + xz + z 2 = 4. (5 marks)
∂x
(c) Find the critical points of the function f (x, y) = x3 − y 2 − xy + 1 and identify the nature of each
point. (11 marks)
 
∂z 1 1
ans: (b) = −9/4. (c) (0, 0)saddle, − , minimum.
∂x 6 12

Page 7 of 7