PARTIAL DIFFERENTIATION AND

ITS APPLICATIONS
´ Function of Several variables – Introduction

´ LIMITS:
The function z = 𝑓 𝑥, 𝑦 ∶ 𝑅×𝑅 → 𝑅 is said to tend to the limit 𝑙 as (𝑥, 𝑦) →
(𝑎,b) if for any 𝜖 > 0, there exists 𝛿 > 0 such that
𝑓 𝑥, 𝑦 − 𝑙 < 𝜖 whenever 𝑥, 𝑦 − (𝑎, 𝑏) <𝛿.

Equivalently, 𝑙 − 𝜖 < 𝑓 𝑥, 𝑦 < 𝑙 + 𝜖 whenever 𝑥 − 𝑎 !+ 𝑦 − 𝑏 ! < 𝛿.

Equivalently, 𝑙 − 𝜖 < 𝑓 𝑥, 𝑦 < 𝑙 + 𝜖 whenever max 𝑥 − 𝑎 , 𝑦 − 𝑏 < 𝛿.

Equivalently, 𝑙 − 𝜖 < 𝑓 𝑥, 𝑦 < 𝑙 + 𝜖 whenever a − 𝛿 < 𝑥 < 𝑎 + 𝛿, b − 𝛿 < 𝑦 < 𝑏 + 𝛿.

´ LIMITS:
The function z = 𝑓 𝑥, 𝑦 is said to tend to the limit 𝑙 as 𝑥 → 𝑎 and 𝑦 → 𝑏 if
and only if the limit 𝑙 is independent of the path followed by the points (𝑥, 𝑦)
as 𝑥 → 𝑎 and 𝑦 → 𝑏 and we write
lim 𝑓 𝑥, 𝑦 = 𝑙
"→$
%→&

Note: In function of one variable, we have only two paths to approach a

point. But in function of two variables, we have infinite path to approach a
point
´ CONTINUITY:
A function z = 𝑓 𝑥, 𝑦 is said to be continuous at the point (𝑎, 𝑏) if
lim 𝑓 𝑥, 𝑦 exists and = 𝑓(𝑎, 𝑏)
"→$
%→&

´ NOTE:
In most of the cases, the limit is the same irrespective of the path along
which the point 𝑥, 𝑦 approaches 𝑎, 𝑏 and
lim lim 𝑓(𝑥, 𝑦) = lim lim 𝑓(𝑥, 𝑦)
"→$ %→& %→& "→$
But it is not always so, as the following examples show:

"'%
lim as (𝑥, 𝑦) → (0,0) along the line 𝑦 = 𝑚𝑥
"(%
"'*" +'*
= lim = , which is different for lines with different slopes.
"→) "(*" +(*

Also,
𝒙'𝒚 𝒙 𝒙'𝒚 '𝒚
𝒍𝒊𝒎 𝒍𝒊𝒎 = 𝒍𝒊𝒎 = 𝟏, whereas 𝒍𝒊𝒎 𝒍𝒊𝒎 = 𝒍𝒊𝒎 = −𝟏.
𝒙→𝟎 𝒚→𝟎 𝒙(𝒚 𝒙→𝟎 𝒙 𝒚→𝟎 𝒙→𝟎 𝒙(𝒚 𝒚→𝟎 𝒚

As (𝑥, 𝑦) is made to approach 0,0 along different paths, 𝑓(𝑥, 𝑦) approaches

different limits. Hence the two repeated limits are not equal and 𝑓(𝑥, 𝑦) is
discontinuous at the origin.
)
Also the function is not defined at (0,0) since 𝑓 𝑥, 𝑦 = for
)
𝑥 = 0, 𝑦 = 0.
´ RESULTS:
(1) If "→$
lim 𝑓 𝑥, 𝑦 = 𝑙 and "→$
lim 𝑔 𝑥, 𝑦 = 𝑚, then
%→& %→&

(a) "→$
lim [𝑓 𝑥, 𝑦 ± 𝑔(𝑥, 𝑦)] = 𝑙 ± 𝑚
%→&

(b) "→$
lim 𝑓 𝑥, 𝑦 · 𝑔 𝑥, 𝑦 =𝑙·𝑚
%→&

(c) "→$
lim 𝑓 𝑥, 𝑦 /𝑔 𝑥, 𝑦 = 𝑙/𝑚 (𝑚 ≠ 0, 𝑔 𝑥, 𝑦 ≠ 0)
%→&

(2) If 𝑓 𝑥, 𝑦 , 𝑔 𝑥, 𝑦 are continuous at 𝑎, 𝑏 then so also are the functions

𝑓 𝑥, 𝑦 ± 𝑔 𝑥, 𝑦 , 𝑓 𝑥, 𝑦 · 𝑔 𝑥, 𝑦 and 𝑓(𝑥, 𝑦)/𝑔 𝑥, 𝑦 provided 𝑔 𝑥, 𝑦 ≠ 0 in the
last case.
PROBLEMS:
1) Evaluate the following limits:
"(%'+)
(a) lim .
"→+ %("'+)
%→+

(Does not exist)

"%(+
(b)"→1
lim . (Take 𝑦 = 𝑓(𝑥))
"! (!%!
%→!

(zero)
"'%
2) If 𝑓 𝑥, 𝑦 = , show that lim lim 𝑓(𝑥, 𝑦) ≠ lim lim 𝑓(𝑥, 𝑦)
!"(% "→) %→) %→) "→)

Also show that the function is discontinuous at the origin.

PARTIAL DERIVATIVES:
Let 𝑧 = 𝑓 𝑥, 𝑦 be a function of two variables 𝑥 and 𝑦.
The derivative of 𝑧 with respect to 𝑥 treating 𝑦 as constant , is called the
23
partial derivative of 𝑧 with respect to 𝑥 and is denoted by .
2"
Thus

Similarly,

23
Note: What is the geometrical meaning of ?
2"
In most of the Engineering applications,

But in general they are not equal.

 "% "! '%!
2! 4 2! 4 𝑖𝑓 (𝑥, 𝑦) ≠ (0,0)
Prove ≠ given 𝑓 𝑥, 𝑦 = K "!(%!
2"2% 2%2"
0 𝑖𝑓 𝑥, 𝑦 = (0,0)
2! 4 2! 4
0,0 = −1; 0,0 = 1
2"2% 2%2"

PROBLEMS:
1) Find the first and second partial derivatives of 𝑧 = 𝑥 5 + 𝑦 5 − 3𝑎𝑥𝑦.
Solution:
23
= 3𝑥 ! − 3𝑎𝑦
2"
23
= 3𝑦 ! − 3𝑎𝑥
2%
2! 3
= 6𝑥
2"!
2! 3
= 6𝑦
2%!
2! 3 2! 3
Also observe that = .
2%2" 2"2%
 2! 3 2 3!
2) If 𝑧 = 𝑓 𝑥 + 𝑐𝑡 + 𝜑 𝑥 − 𝑐𝑡 , prove that 26 !
= 𝑐 ! 2"!.
This is an important partial differential equation, known as wave equation.

2! 8 2! 8 2! 8
3) If 𝑣 = (𝑥 ! + 𝑦! + 𝑧 ! )'+/! , prove that + + = 0.
2"! 2%! 23 !

This equation is known as Laplace equation.

A function 𝑣 satisfying the Laplace equation is said to be a harmonic function.

2 2 2 ! '9
4) If 𝑢 = log 𝑥 5 + 𝑦 5 + 𝑧 5 − 3𝑥𝑦𝑧 , show that ( + + ) 𝑢 = .
2" 2% 23 ("(%(3)!

5) Find the value of 𝑛 so that the equation 𝑉 = 𝑟 : (3𝑐𝑜𝑠 ! 𝜃 − 1) satisfies the

2 2< + 2% 2<
relation 2; 𝑟 ! 2; + =>:? 2? 𝑠𝑖𝑛𝜃 2? = 0.
(𝑛 = 2, −3)
HOMOGENEOUS FUNCTIONS:
An expression of the form 𝑎) 𝑥 : + 𝑎+ 𝑥 :'+ 𝑦 + 𝑎! 𝑥 :'! 𝑦 ! + ⋯ + 𝑎: 𝑦 : in which
every term is of the 𝑛th degree, is called a homogeneous function of degree
𝑛.

%
Any function 𝑓 𝑥, 𝑦 which can be expressed in the form 𝑥 : 𝜑 , is called a
"
homogeneous function of degree 𝑛 in 𝑥 and 𝑦.
Give an example of an homogeneous function of degree −1.

EULER’S THEOREM ON HOMOGENEOUS FUNCTIONS:

If 𝑢 be a homogeneous function of degree 𝑛 in 𝑥 and 𝑦, then
PROBLEMS:
" " %" 3 " "%(%3(3" 2@ 2@ 2@
1) If 𝑢 = + 𝑙𝑜𝑔 , find the value of 𝑥 + 𝑦 + 𝑧 .
"" (%" (3 " "! (%! (3 ! 2" 2% 23
2@ 2@ 2@ " " %" 3 "
(𝑥 + 𝑦 + 𝑧 = 6 " " ")
2" 2% 23 " (% (3

2) If 𝑧 is a homogeneous function of degree 𝑛 in 𝑥 and 𝑦, show that

𝜕 !𝑧 𝜕 !𝑧 𝜕 !𝑧
𝑥 ! ! + 2xy ! + 𝑦 ! ! = n n − 1 z.
𝜕𝑥 𝜕𝑦 𝜕𝑧
3) Verify Euler’s theorem when 𝑓 𝑥, 𝑦 = 3𝑥 ! 𝑦𝑧 + 5𝑥𝑦 ! 𝑧 + 4𝑧 A .
" (% " " 2@ 2@
4)If 𝑢 = tan'+ , prove that 𝑥 + 𝑦 = sin 2𝑢 and
"(% 2" 2%
!
2 @ 2! @ 2 @ !
𝑥! ! + 2xy + 𝑦! ! = 2 cos 3𝑢 sin 𝑢.
2" 2"2% 2%
% "
5)Given 𝑧 = 𝑥 : 𝑓+ + 𝑦 ': 𝑓! , prove that
" %
2 3 ! 2! 3 2! 3 23 23
𝑥! ! + 2xy + 𝑦 ! +𝑥 +𝑦 = 𝑛! 𝑧.
2" 2"2% 2%! 2" 2%
TOTAL DERIVATIVE:
If 𝑢 = 𝑓 𝑥, 𝑦 , where 𝑥 = 𝜑(𝑡) and 𝑦 = Ψ 𝑡 , then we can express 𝑢 as a
function of 𝑡 alone by substituting the values of 𝑥 and 𝑦 in 𝑓 𝑥, 𝑦 .
B@
Thus we can find the ordinary derivative which is called the total
B6
2@ 2@
derivative of 𝑢 to distinguish it from the partial derivatives and .
2" 2%

CHAIN RULE:

COROLLARY:

Taking 𝑡 = 𝑥, chain rule becomes

DIFFERENTIATION OF IMPLICIT FUNCTIONS:
If 𝑓 𝑥, 𝑦 = c be an implicit relation between 𝑥 and 𝑦 which defines as a
differentiable function of 𝑥, then

This gives the important formula

for the first differential coefficient of an implicit function.

CHANGE OF VARIABLES:
If u = 𝑓 𝑥, 𝑦 where 𝑥 = 𝜑(𝑠, 𝑡) and 𝑦 = Ψ 𝑠, 𝑡 , then we have

where the ordinary derivatives have been replaced by the partial derivatives
because 𝑥, 𝑦 are functions of two variables 𝑠 and 𝑡.
Similarly, regarding 𝑠 as constant, we obtain
PROBLEMS:
B! % C! ;'!DC=(D! =
1) If 𝑓 𝑥, 𝑦 = 0, show that = − .
B"! C"
B@
2) If 𝑢 = 𝑥 ! + 𝑦 ! + 𝑧 ! and 𝑥 = 𝑒 !6 , 𝑦 = 𝑒 !6 cos 3𝑡 , 𝑧 = 𝑒 !6 sin 3𝑡. Find as a
B6
total derivative and verify the result by direct substitution.
3) If the curves 𝑓 𝑥, 𝑦 = 0 and φ 𝑥, 𝑦 = 0 touch, show that at the point
24 2E 24 2E
of contact · = ·
2" 2% 2% 2"
2! @ 2! @
4) Transform the equation + ! = 0 into polar coordinates.
2"! 2%
2! @ + 2@ + 2! @
(The transformed equation is + + ! ! = 0.)
2; ! ; 2; ; 2?
5) If by the substitution 𝑢 = 𝑥 ! − 𝑦 ! , 𝑣 = 2𝑥𝑦, 𝑓 𝑥, 𝑦 = 𝜃 𝑢, 𝑣 , show that
2! 4 2! 4 2! ? 2! ?
+ = 4(𝑥 ! + !
𝑦 )( ! + ).
2"! 2%! 2@ 28!
JACOBIANS:
If 𝑢 and 𝑣 are functions of two independent variables 𝑥 and 𝑦, then the
determinant is called the Jacobian of 𝑢, 𝑣 with

respect to 𝑥, 𝑦 , and is written as or .

Similarly, the Jacobian of 𝑢, 𝑣, 𝑤 with respect to 𝑥, 𝑦, 𝑧 is

PROPERTIES OF JACOBIANS:
2(@,8) 2(",%)
1) If 𝐽 = and 𝐽G = then 𝐽𝐽G = 1.
2(",%) 2(@,8)

2) Chain rule for Jacobians:

If 𝑢, 𝑣 are functions of 𝑟, 𝑠 and 𝑟, 𝑠 are functions of 𝑥, 𝑦, then

PROBLEMS:
2(",%)
(1)In polar coordinates 𝑥 = 𝑟𝑐𝑜𝑠𝜃, 𝑦 = 𝑟𝑠𝑖𝑛𝜃, show that = 𝑟.
2(;,?)

(2)In cylindrical coordinates, 𝑥 = 𝜌𝑐𝑜𝑠𝜑, 𝑦 = 𝜌𝑠𝑖𝑛𝜑, 𝑧 = 𝑧, show that

2(",%,3)
= 𝜌.
2(H,E,3)

(3)In spherical coordinates, 𝑥 = 𝑟𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜑, 𝑦 = 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜑, 𝑧 = 𝑟 𝑐𝑜𝑠𝜃, show

2(",%,3)
that = 𝑟 ! sin 𝜃.
2(;,?,E)
JACOBIAN OF IMPLICIT FUNCTIONS:
If 𝑢+ , 𝑢! , 𝑢5 instead of being given explicitly in terms 𝑥+ , 𝑥! , 𝑥5 be connected
with them equations such as
𝑓+ (𝑢+ , 𝑢! , 𝑢5 , 𝑥+ , 𝑥! , 𝑥5 ) = 0,
𝑓! (𝑢+ , 𝑢! , 𝑢5 , 𝑥+ , 𝑥! , 𝑥5 ) = 0,
𝑓5 (𝑢+ , 𝑢! , 𝑢5 , 𝑥+ , 𝑥! , 𝑥5 ) = 0,

then

FUNCTIONAL RELATIONSHIP:
If 𝑢+ , 𝑢! , 𝑢5 be functions of 𝑥+ , 𝑥! , 𝑥5 then the necessary and sufficient
condition for the existence of a functional relationship of the form
𝑓 𝑢+ , 𝑢! , 𝑢5 = 0, is
PROBLEMS:
2(",%,3)
1) If 𝑢 = 𝑥𝑦𝑧, 𝑣 = 𝑥 ! + 𝑦 ! + 𝑧 ! , 𝑤 = 𝑥 + 𝑦 + 𝑧, 𝑓𝑖𝑛𝑑 .
2(@,8,I)
2(",%,3)
( = 1/2(𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥))
2(@,8,I)

2) If 𝑢 = 𝑥 (1 − 𝑦 ! ) + 𝑦 (1 − 𝑥 ! , v = sin'+ 𝑥 + sin'+ 𝑦 , show that 𝑢, 𝑣 are

functionally related and find the relationship.

(𝑢 = sin 𝑣 is the required relationship between 𝑢 and 𝑣).

MAXIMA AND MINIMA OF FUNCTONS OF TWO VARIABLES:
A function 𝑓 𝑥, 𝑦 is said to have a maximum or minimum at 𝑥 = 𝑎, 𝑦 = 𝑏,
according as

for all positive or negative small values of ℎ and 𝑘.

In other words, if is of the same sign for all small

values of ℎ, 𝑘, and if this sign is negative, then 𝑓(𝑎, 𝑏) is a maximum. If this
sign is positive, 𝑓(𝑎, 𝑏) is a minimum.
Considering 𝑧 = 𝑓(𝑥, 𝑦) as a surface, maximum value of 𝑧 occurs at the top of
an elevation from which the surface descends in every direction and a
minimum value occurs at the bottom of a depression from which the surface
ascends in every direction.

Sometimes the maximum or minimum value may for a ridge such that the
surface descends or ascends n all directions except that of the ridge.
Besides these, we have such a point of the surface, where the tangent plane
is horizontal and the surface looks like leather seat on the horse’s back which
falls for displacement in certain directions and rises for displacements in
other directions. Such a point is called a saddle point.

NOTE: A maximum or minimum value of a function is called its extreme

value.
CONDITIONS FOR 𝑓 𝑥, 𝑦 TO BE A MAXIMUM OR MINIMUM:
By Taylor’s theorem, we have

For small values of ℎ and 𝑘, the second and higher order terms are still smaller
and hence they neglected. Thus

Taking ℎ = 0 we see that the right hand side changes sign when 𝑘 changes
sign. Hence 𝑓 𝑥, 𝑦 cannot have a maximum or minimum at (𝑎, 𝑏) unless
𝑓% 𝑎, 𝑏 = 0.
Similarly, taking k= 0 we find that 𝑓 𝑥, 𝑦 cannot have a maximum or minimum
at 𝑎, 𝑏 unless 𝑓" 𝑎, 𝑏 = 0.
Hence the necessary conditions for 𝑓 𝑥, 𝑦 to have a maximum or minimum at
(𝑎, 𝑏) are that
For small values of ℎ and 𝑘,
where
Thus

If 𝑟𝑡 − 𝑠 ! > 0, then 𝑓(𝑥, 𝑦) has a maximum or minimum at 𝑎, 𝑏 according as

𝑟 < 𝑜𝑟 > 0.
If 𝑟𝑡 − 𝑠 ! < 0, then ∆ will change with ℎ and 𝑘 and hence there is no
maximum or minimum at (𝑎, 𝑏) i.e., it is a saddle point.
If 𝑟𝑡 − 𝑠 ! = 0, further investigation is required to find whether there is a
maximum or minimum at 𝑎, 𝑏 or not.

STATIONARY VALUE:
𝑓(𝑎, 𝑏) is said to be a stationary value of 𝑓(𝑥, 𝑦) if 𝑓" 𝑎, 𝑏 = 0 and 𝑓% 𝑎, 𝑏 = 0
i.e., the function is stationary at 𝑎, 𝑏 .
Working rule to find the maximum and minimum values of 𝑓(𝑥, 𝑦)
24 24
1) Find and and equate each to zero. Solve these as simultaneous
2" 2%
equations in 𝑥 and 𝑦. Let 𝑎, 𝑏 , 𝑐, 𝑑 , … be the pairs of values.

2! 4 2! 4 2! 4
2) Calculate the value of 𝑟 = , 𝑠= , 𝑡= for each pair of values.
2"! 2"2% 2%!

3) (i) If 𝑟𝑡 − 𝑠 ! > 0 and 𝑟 < 0 at 𝑎, 𝑏 , 𝑓(𝑎, 𝑏) is a maximum value.

(ii) If 𝑟𝑡 − 𝑠 ! > 0 and 𝑟 > 0 𝑎, 𝑏 , 𝑓(𝑎, 𝑏) is a minimum value.
(iii) If 𝑟𝑡 − 𝑠 ! < 0 𝑎𝑡 𝑎, 𝑏 , 𝑓(𝑎, 𝑏) is not an extreme value, i.e., 𝑎, 𝑏 is a
saddle point.
(iv) If 𝑟𝑡 − 𝑠 ! = 0 𝑎𝑡 𝑎, 𝑏 , the case is doubtful and needs further
investigation.
Similarly examine the other pairs of values one by one.
PROBLEMS:
1) Discuss the maxima and minima of 𝒇 𝒙, 𝒚 = 𝒙𝟑 𝒚𝟐 (1-x-y).
+ + + +
The stationary points are , , 0,0 . At , , 𝑓 𝑥, 𝑦 has a maximum and
! 5 ! 5
+
maximum value = .
A5!
At 0,0 , 𝑟𝑡 − 𝑠 ! = 0 and therefore further investigation is needed.
For points along the line 𝑦 = 𝑥, 𝑓 𝑥, 𝑦 = 𝑥 L 1 − 2𝑥 which is positive for 𝑥 = 0.1
i.e., in the neighbourhood of (0,0) there are points where 𝑓 𝑥, 𝑦 > 𝑓(0,0) and
there are points where 𝑓 𝑥, 𝑦 < 𝑓(0,0). Hence 𝑓 0,0 is not an extreme value.
2) In a plane triangle, find the maximum value of 𝐜𝐨𝐬 𝑨 𝐜𝐨𝐬 𝑩 𝐜𝐨𝐬 𝑪.
𝑓 𝐴, 𝐵 = − cos 𝐴 cos 𝐵 cos(𝐴 + 𝐵)
M + 5
𝐴 = 𝐵 = , r = −1, s = − , t = −1 so that 𝑟𝑡 − 𝑠 ! = .
5 ! A
M M
These show that 𝑓 𝐴, 𝐵 is maximum for 𝐴 = 𝐵 = . Then C = . Hence
5 5
M
cos 𝐴 cos 𝐵 cos 𝐶 is maximum when each of the angles is 5
i.e., triangle is
equilateral and its maximum value = 1/8.