Multivariate Calculus: MTH-174 Prof. U.

Sarkar

Limit
Definition: A real number 𝑙 is called a limit (functional value) of a function 𝑓(𝑥, 𝑦) of two variables 𝑥
and 𝑦 at a point (𝑎, 𝑏) if
lim 𝑓(𝑥, 𝑦) = 𝑙
𝑥 ,𝑦 →(𝑎,𝑏)
Special case: Limit of a function 𝑓(𝑥, 𝑦) at the origin-
𝜙 (𝑥,𝑦 )
The technique for choosing the general path is applicable when 𝑓 𝑥, 𝑦 = 𝑇 𝑝 ±𝑇𝑞
1 2
𝑝
(i) If 𝑝 ≥ 𝑞, compute 𝑛 = 𝑞 and set the relation as 𝑇2 = 𝑚𝑇1𝑛 and assume 𝑥, 𝑦 → (0,0) along

this path.
𝑞
(ii) If 𝑞 ≥ 𝑝, compute 𝑛 = 𝑝 and set the relation as 𝑇1 = 𝑚𝑇2𝑛 and assume 𝑥, 𝑦 → (0,0) along

this path.
If, along this general path, we get a unique finite value then the limit exists. Otherwise we say the limit
does not exist.
Continuity: A function 𝑓(𝑥, 𝑦) is continuous at a point (𝑎, 𝑏) if
lim 𝑓(𝑥, 𝑦) = 𝑓(𝑎, 𝑏)
𝑥 ,𝑦 →(𝑎,𝑏)
Example 1: Prove that the following limit does not exist.
𝑥−𝑦
lim
𝑥 ,𝑦 →(0,0) 𝑥+𝑦

Solution: We have
𝑥−𝑦
lim
𝑥,𝑦 →(0,0) 𝑥+𝑦

Let 𝑥, 𝑦 → (0,0) along the straight line 𝑦 = 𝑚𝑥

Then along 𝑦 = 𝑚𝑥
𝑥−𝑦 𝑥−𝑚𝑥 𝑥 (1−𝑚 ) 1−𝑚 1−𝑚
lim = lim 𝑥+𝑚𝑥 = lim 𝑥 (1+𝑚 ) = lim 1+𝑚 = 1+𝑚
𝑥,𝑦 →(0,0) 𝑥+𝑦 𝑥 →0 𝑥→0 𝑥→0

This value is not unique as 𝑚 can take any real value. Hence the limit does not exist.
Example 2: Prove that the following function is not continuous at the origin.

𝑥𝑦3
𝑥 2 +𝑦 6
, 𝑓𝑜𝑟 𝑥, 𝑦 ≠ (0,0)
𝑓 𝑥, 𝑦 =
0 , 𝑓𝑜𝑟 𝑥, 𝑦 = (0,0)

Solution: Clearly the function is continuous for all (𝑥, 𝑦) when 𝑥, 𝑦 ≠ (0,0). We have to check

lim 𝑓(𝑥, 𝑦) = 𝑓(0,0)

𝑥,𝑦 → 0,0

𝑥𝑦3
Now lim 𝑓 𝑥, 𝑦 = lim
𝑥,𝑦 →(0,0) 𝑥 ,𝑦 →(0,0) 𝑥 2 +𝑦 6

Let 𝑥, 𝑦 → (0,0) along the path 𝑥 = 𝑚𝑦 3 .

 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Then along 𝑥 = 𝑚𝑦 3

𝑥𝑦 3 𝑚𝑦 3 . 𝑦 3
lim 𝑓 𝑥, 𝑦 = lim = lim
𝑥,𝑦 → 0,0 𝑥 ,𝑦 → 0,0 𝑥 2 + 𝑦 6 𝑦 →0 𝑚𝑦 3 2 + 𝑦 6

𝑦6𝑚 𝑚 𝑚
= lim 6 2
= lim 2
= 2
𝑦 →0 𝑦 (𝑚 + 1) 𝑦 →0 (𝑚 + 1) 𝑚 +1

This value is not unique as 𝑚 can take any real value. Hence the limit does not exist and therefore, it is
not continuous at the origin.

Example 3: Test the continuity of the following function at the origin.

𝑥 2𝑦 2
, 𝑓𝑜𝑟 𝑥, 𝑦 ≠ (0,0)
𝑓 𝑥, 𝑦 = 𝑥 2 𝑦 2 +(𝑥−𝑦 )2
0 , 𝑓𝑜𝑟 𝑥, 𝑦 = (0,0)

Solution: Clearly the function is continuous for all (𝑥, 𝑦) when 𝑥, 𝑦 ≠ (0,0). We have to check

lim 𝑓(𝑥, 𝑦) = 𝑓(0,0)

𝑥,𝑦 → 0,0

𝑥 2𝑦 2
Now lim 𝑓 𝑥, 𝑦 = lim
𝑥,𝑦 →(0,0) 𝑥 ,𝑦 →(0,0) 𝑥 2 𝑦 2 +(𝑥−𝑦 )2

Let 𝑥, 𝑦 → (0,0) along the path 𝑥 − 𝑦 = 𝑚𝑥𝑦. Then along 𝑥 − 𝑦 = 𝑚𝑥𝑦

𝑥 2𝑦 2 𝑥 2𝑦 2 1 1
lim 𝑓(𝑥, 𝑦) = lim = lim = lim = 𝑚 2 +1
𝑥,𝑦 → 0,0 𝑥,𝑦 →(0,0) 𝑥 2 𝑦 2 +𝑚 2 𝑥 2 𝑦 2 𝑥,𝑦 →(0,0) 𝑥 2 𝑦 2 (1+𝑚 2 ) 𝑥 ,𝑦 →(0,0) (𝑚 2 +1)

This value is not unique as 𝑚 can take any real value. Hence the limit does not exist and therefore, it is
not continuous at the origin.

Example 4: Prove the following.

𝑥−𝑦 𝑥 −𝑦
lim lim ≠ lim lim
𝑥 →0 𝑦 →0 𝑥+𝑦 𝑦 →0 𝑥 →0 𝑥 +𝑦

Solution: We have
𝑥 −𝑦 𝑥 −0 𝑥 𝑥−𝑦 0−𝑦 −𝑦
lim lim 𝑥 +𝑦 = lim 𝑥 +0 = lim 𝑥 = 1 and lim lim 𝑥+𝑦 = lim 0+𝑦 = lim = −1
𝑥 →0 𝑦 →0 𝑥→0 𝑥→0 𝑦 →0 𝑥→0 𝑦→0 𝑦 →0 𝑦

Clearly, the limits are not equal.

Example 5: Show that the following limits are equal only when 𝑎𝑑 = 𝑏𝑐.
𝑎𝑥 +𝑏𝑦 𝑎𝑥 +𝑏𝑦
lim lim 𝑐𝑥 +𝑑𝑦 & lim lim 𝑐𝑥 +𝑑𝑦
𝑥 →0 𝑦 →0 𝑦 →0 𝑥→0
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Solution: We have
𝑎𝑥 +𝑏𝑦 𝑎𝑥 −𝑏×0 𝑎𝑥 𝑎 𝑎𝑥 +𝑏𝑦 𝑎 ×0+𝑏𝑦 𝑏𝑦 𝑏
lim lim 𝑐𝑥 +𝑑𝑦 = lim 𝑐𝑥 +𝑑×0 = lim 𝑐𝑥 = 𝑐
and lim lim 𝑐𝑥 +𝑑𝑦 = lim 𝑐×0+𝑑𝑦 = lim 𝑑𝑦 = 𝑑
𝑥 →0 𝑦 →0 𝑥 →0 𝑥 →0 𝑦 →0 𝑥→0 𝑦 →0 𝑦 →0

Clearly, the limits are equal if

𝑎 𝑏
= ⇒ 𝑎𝑑 = 𝑏𝑐
𝑐 𝑑

Problem Set-I

1. Show that lim 𝑓 𝑥, 𝑦 does not exist when

𝑥,𝑦 →(0,0)

𝑥 2 +5𝑦 2
(i) 𝑓 𝑥, 𝑦 = (𝑥−𝑦 )2
𝑥𝑦 (𝑥 −2𝑦 )
(ii) 𝑓 𝑥, 𝑦 = 𝑥 3 +𝑦 3

2. Test the continuity of the following functions at the origin

𝑥 2𝑦
𝑥 4 +𝑦 2
, 𝑓𝑜𝑟 𝑥, 𝑦 ≠ (0,0)
(i) 𝑓 𝑥, 𝑦 =
0 , 𝑓𝑜𝑟 𝑥, 𝑦 = (0,0)

Partial Derivatives
The derivative of a function of several independent variables w.r.t. any one, keeping the others constant.

𝜕𝑢 𝜕𝑓 𝜕𝑢 𝜕𝑓 𝜕𝑢 𝜕𝑓
Let 𝑢 = 𝑓(𝑥, 𝑦, 𝑧). Then = , = , = are respectively called the 1st order partial derivatives
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑧 𝜕𝑧
𝜕𝑓 𝑓 𝑥+𝑕 ,𝑦 ,𝑧 −𝑓(𝑥,𝑦 ,𝑧)
w.r.t. 𝑥, 𝑦 and 𝑧 where 𝜕𝑥 = lim 𝑕
and so on.
𝑕 →0

𝜕2𝑢 𝜕2𝑢 𝜕2𝑢 𝜕2𝑢

Similarly, if 𝑢 = 𝑓(𝑥, 𝑦) then 𝜕 𝑥 2 , 𝜕𝑥𝜕𝑦 , 𝜕𝑦𝜕𝑥 , 𝜕𝑦 2 are the 2nd order partial derivatives. Note that it is not
𝜕2𝑢 𝜕2𝑢
in general true that = .
𝜕𝑥𝜕𝑦 𝜕𝑦𝜕𝑥

𝜕2𝑢 𝜕2𝑢 𝑦
Example 6: Verify that = if 𝑢 = tan−1 .
𝜕𝑥𝜕𝑦 𝜕𝑦𝜕 𝑥 𝑥

Solution: We have
𝑦
𝑢 = tan−1 𝑥

Therefore,

𝜕𝑢 1 𝑦 −𝑦 𝜕2𝑢 𝑥 2 +𝑦 2 ×1−𝑦 ×2𝑦 𝑦 2 −𝑥 2

𝜕𝑥
= 𝑦 2
× − 𝑥 2 = 𝑥 2 +𝑦 2 and hence 𝜕𝑦𝜕𝑥 = − 𝑥 2 +𝑦 2 2
= 𝑥 2 +𝑦 2 2
1+
𝑥
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝜕𝑢 1 1 𝑥 𝜕2𝑢 𝑥 2 +𝑦 2 ×1−𝑥 ×2𝑥 𝑦 2 −𝑥 2

𝜕𝑦
= 𝑦 2
× 𝑥
= 𝑥 2 +𝑦 2 and hence 𝜕𝑥𝜕𝑦 = 𝑥 2 +𝑦 2 2
= 𝑥 2 +𝑦 2 2
1+
𝑥

𝜕2𝑢 𝜕2𝑢
Thus, 𝜕𝑥𝜕𝑦 = 𝜕𝑦𝜕𝑥 .

𝜕2𝑢 𝜕2𝑢 𝑦 𝑥
Example 7: Verify that = if 𝑢 = 𝑥 2 tan−1 − 𝑦 2 tan−1 .
𝜕𝑥𝜕𝑦 𝜕𝑦𝜕𝑥 𝑥 𝑦

Solution: Given
𝑦 𝑥
𝑢 = 𝑥 2 tan−1 − 𝑦 2 tan−1
𝑥 𝑦

Therefore,
𝜕𝑢 𝑦 1 𝑦 1 1
𝜕𝑥
= 2𝑥 tan−1 𝑥
+ 𝑥2 × 𝑦 2
× − 𝑥 2 − 𝑦2 × 𝑥 2
× 𝑦
1+ 1+
𝑥 𝑦
𝑦 𝑥 2𝑦 𝑦3
= 2𝑥 tan−1 − −
𝑥 𝑥 2 +𝑦 2 𝑥 2 +𝑦 2
−1 𝑦
= 2𝑥 tan 𝑥
−𝑦
𝜕𝑢 𝑥
Similarly, = 𝑥 − 2𝑦 tan−1
𝜕𝑦 𝑦
𝜕2𝑢 1 1 𝑥 2 −𝑦 2
∴ 𝜕𝑦𝜕𝑥 = 2𝑥 × 𝑦 2
× 𝑥
− 1 = 𝑥 2 +𝑦 2 and
1+
𝑥
𝜕2𝑢 1 1 𝑥 2 −𝑦 2
𝜕𝑥𝜕𝑦
= 1 − 2𝑦 × 𝑥 2
× 𝑦
= 𝑥 2 +𝑦 2
1+
𝑦
𝜕2𝑢 𝜕2𝑢
Thus, 𝜕𝑥𝜕𝑦 = 𝜕𝑦𝜕𝑥 .
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
Example 8: If 𝑢 = 𝑥 2 + 𝑦 2 + 𝑧 2 −1/2
, show that 𝜕 𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2 = 0.

Solution: Given
𝑢 = 𝑥2 + 𝑦 2 + 𝑧2 −1/2

𝜕𝑢 1
𝜕𝑥
= − 2 𝑥2 + 𝑦 2 + 𝑧2 −3/2
× 2𝑥 = −𝑥 𝑥 2 + 𝑦 2 + 𝑧 2 −3/2

3 5
𝜕2𝑢 − 3 −
∴ = − 𝑥2 + 𝑦 2 + 𝑧2 2 −𝑥 − 𝑥2 + 𝑦 2 + 𝑧2 2 × 2𝑥
𝜕𝑥2 2
2 2 2 −5/2 2 2 2
= 𝑥 +𝑦 +𝑧 2𝑥 − 𝑦 − 𝑧
Similarly,
𝜕2𝑢
𝜕𝑦 2
= 𝑥2 + 𝑦 2 + 𝑧2 −5/2
2𝑦 2 − 𝑧 2 − 𝑥 2 and
𝜕2𝑢
𝜕𝑧2
= 𝑥2 + 𝑦 2 + 𝑧2 −5/2
2𝑧 2 − 𝑥 2 − 𝑦 2
Therefore,
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
𝜕𝑥2
+ 𝜕𝑦 2 + 𝜕 𝑧 2 = 𝑥 2 + 𝑦 2 + 𝑧 2 −5/2
2𝑥 2 − 𝑦 2 − 𝑧 2 + 2𝑦 2 − 𝑧 2 − 𝑥 2 + 2𝑧 2 − 𝑥 2 − 𝑦 2
= 𝑥2 + 𝑦 2 + 𝑧2 −5/2
×0=0

Example 9: If 𝑢 = log 𝑥 3 + 𝑦 3 + 𝑧 3 − 3𝑥𝑦𝑧 , show that

𝜕𝑢 𝜕𝑢 𝜕𝑢 3
(i) 𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 = 𝑥 +𝑦 +𝑧
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝜕 𝜕 𝜕 2 9
(ii) 𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 𝑢 = − (𝑥+𝑦 +𝑧)2

Solution: Given
𝑢 = log 𝑥 3 + 𝑦 3 + 𝑧 3 − 3𝑥𝑦𝑧 = log{ 𝑥 + 𝑦 + 𝑧 𝑥 + 𝑦𝜔 + 𝑧𝜔2 (𝑥 + 𝑦𝜔2 + 𝑧𝜔)}
= log 𝑥 + 𝑦 + 𝑧 + log 𝑥 + 𝑦𝜔 + 𝑧𝜔2 + log(𝑥 + 𝑦𝜔2 + 𝑧𝜔)
𝜕𝑢 1 1 1
∴ 𝜕𝑥 = 𝑥+𝑦 +𝑧 + 𝑥+𝑦𝜔 +𝑧𝜔 2 + 𝑥+𝑦 𝜔 2 +𝑧𝜔
𝜕𝑢 1 𝜔 𝜔2
𝜕𝑦
= 𝑥+𝑦 +𝑧 + 𝑥+𝑦𝜔 +𝑧𝜔 2 + 𝑥 +𝑦 𝜔 2 +𝑧𝜔
𝜕𝑢 1 𝜔2 𝜔
= + +
𝜕𝑧 𝑥+𝑦 +𝑧 𝑥+𝑦𝜔 +𝑧𝜔 2 𝑥 +𝑦 𝜔 2 +𝑧𝜔
Therefore,
𝜕𝑢 𝜕𝑢 𝜕𝑢 1+1+1 1+𝜔 +𝜔 2 1+𝜔 +𝜔 2 3
+ + = + + = since 1 + 𝜔 + 𝜔2 = 0.
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝑥 +𝑦 +𝑧 𝑥 +𝑦𝜔 +𝑧𝜔 2 𝑥 +𝑦 𝜔 2 +𝑧𝜔 𝑥+𝑦 +𝑧
Now,
𝜕 𝜕 𝜕 2 𝜕 𝜕 𝜕 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕 𝜕 𝜕 3
𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 𝑢= 𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 = 𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 𝑥+𝑦 +𝑧
𝜕 3 𝜕 3 𝜕 3
= 𝜕𝑥 𝑥+𝑦 +𝑧
+ 𝜕𝑦 𝑥+𝑦 +𝑧
+ 𝜕𝑧 𝑥+𝑦+𝑧
3 3 3
=− 𝑥+𝑦 +𝑧 2
− 𝑥+𝑦 +𝑧 2
− 𝑥+𝑦 +𝑧 2
9
=−
(𝑥+𝑦 +𝑧)2
𝑥2 𝑦2 𝑧2 𝜕𝑢 2 𝜕𝑢 2 𝜕𝑢 2 𝜕𝑢 𝜕𝑢 𝜕𝑢
Example 10: If + + = 1, prove that + + =2 𝑥 +𝑦 +𝑧 .
𝑎 2 +𝑢 𝑏 2 +𝑢 𝑐 2 +𝑢 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧

Solution: Given that

𝑥2 𝑦2 𝑧2
𝑎 2 +𝑢
+ 𝑏 2 +𝑢 + 𝑐 2 +𝑢 = 1 (1)

Differentiating (1) partially with respect to 𝑥, we get

2𝑥 𝑥2 𝑦2 𝑧2 𝜕𝑢
𝑎 2 +𝑢
− (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢)2 + (𝑐 2 +𝑢)2 𝜕𝑥
=0
−1
𝜕𝑢 2𝑥 𝑥2 𝑦2 𝑧2
⇒ 𝜕𝑥 = 𝑎 2 +𝑢 × (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢)2 + (𝑐 2 +𝑢)2
Similarly,
−1
𝜕𝑢 2𝑦 𝑥2 𝑦2 𝑧2
= × + + and
𝜕𝑦 𝑏 2 +𝑢 (𝑎 2 +𝑢 )2 (𝑏 2 +𝑢)2 (𝑐 2 +𝑢 )2
−1
𝜕𝑢 2𝑧 𝑥2 𝑦2 𝑧2
𝜕𝑧
= 𝑐 2 +𝑢 × (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢)2 + (𝑐 2 +𝑢)2
Therefore,
−1
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝑥2 𝑦2 𝑧2 𝑥2 𝑦2 𝑧2
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 + 𝑧 𝜕𝑧 = 2 𝑎 2 +𝑢
+ 𝑏 2 +𝑢 + 𝑐 2 +𝑢 × 𝑎 2 +𝑢 2 + 𝑏 2 +𝑢 2 + 𝑐 2 +𝑢 2
−1
𝑥2 𝑦2 𝑧2
=2 (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢 )2 + (𝑐 2 +𝑢)2 [using (1)] (2)
Now,
−2
𝜕𝑢 2 4𝑥 2 𝑥2 𝑦2 𝑧2
= 2 × 2 + 2 +
𝜕𝑥 (𝑎 2 +𝑢) (𝑎 2 +𝑢) (𝑏 2 +𝑢) (𝑐 2 +𝑢 )2
Similarly,
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

−2
𝜕𝑢 2 4𝑦 2 𝑥2 𝑦2 𝑧2
𝜕𝑦
= (𝑏 2 +𝑢)2 × (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢)2 + (𝑐 2 +𝑢)2 and
−2
𝜕𝑢 2 4𝑧 2 𝑥2 𝑦2 𝑧2
𝜕𝑧
= (𝑐 2 +𝑢)2 × (𝑎 2 +𝑢)2
+ (𝑏 2 +𝑢)2 + (𝑐 2 +𝑢)2
Therefore,
−2
𝜕𝑢 2 𝜕𝑢 2 𝜕𝑢 2 𝑥2 𝑦2 𝑧2 𝑥2 𝑦2 𝑧2
𝜕𝑥
+ 𝜕𝑦
+ 𝜕𝑧
=4 𝑎 2 +𝑢 2 + 𝑏 2 +𝑢 2 + 𝑐 2 +𝑢 2
× 𝑎 2 +𝑢 2 + 𝑏 2 +𝑢 2 + 𝑐 2 +𝑢 2
−1
𝑥2 𝑦2 𝑧2 𝜕𝑢 𝜕𝑢 𝜕𝑢
=4 𝑎 2 +𝑢 2
+ 𝑏 2 +𝑢 2
+ 𝑐 2 +𝑢 2
= 2 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 + 𝑧 𝜕𝑧 , using (2).

Problem Set-II

𝜕2𝑢 𝜕2𝑢 𝑥 2 +𝑦 2
1. Verify the equality = when 𝑢 = log .
𝜕𝑥𝜕𝑦 𝜕𝑦𝜕𝑥 𝑥𝑦
𝜕𝑢 𝜕𝑢 2 𝜕𝑢 𝜕𝑢
2. If 𝑢 = 𝑥 2 + 𝑦 2 , prove that 𝜕𝑥
− 𝜕𝑦 = 4 1 − 𝜕𝑥 − 𝜕𝑦 .
𝑥 𝑦 𝜕𝑢 𝜕𝑢
3. If 𝑢 = sin−1 𝑦
+ tan−1 𝑥
, prove that 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 0.
4. If 𝑢 = log 𝑟, where 𝑟 2 = (𝑥 − 𝑎)2 + (𝑦 − 𝑏)2 + (𝑧 − 𝑐)2 , show that
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢 1
𝜕𝑥2
+ 𝜕𝑦 2 + 𝜕 𝑧 2 = 𝑟 2
𝜕2𝑧 𝜕2𝑧
5. If 𝑧 = 𝑓 𝑥 + 𝑐𝑡 + 𝑔(𝑥 − 𝑐𝑡), prove that 𝑐 2 𝜕 𝑥 2 = 𝜕 𝑡 2 .
𝜕𝑢 𝜕𝑢 𝜕𝑢
6. If 𝑢 = log tan 𝑥 + tan 𝑦 + tan 𝑧 , prove that sin 2𝑥 + sin 2𝑦 + sin 2𝑧 = 2.
𝜕𝑥 𝜕𝑦 𝜕𝑧

Chain Rules:
𝝏𝒖 𝝏𝒖 𝝏𝒇 𝝏𝒇
 If 𝑢 = 𝑓(𝑥, 𝑦) then 𝒅𝒖 = 𝝏𝒙 𝒅𝒙 + 𝝏𝒚 𝒅𝒚 = 𝝏𝒙 𝒅𝒙 + 𝝏𝒚 𝒅𝒚
𝒅𝒖 𝝏𝒖 𝒅𝒙 𝝏𝒖 𝒅𝒚
 If 𝑢 = 𝑓(𝑥, 𝑦) where 𝑥 = 𝜙(𝑡) and 𝑦 = 𝜓(𝑡) then 𝒅𝒕
= 𝝏𝒙 𝒅𝒕 + 𝝏𝒚 𝒅𝒕
Change of Variables:
 If 𝑢 = 𝑓(𝑥, 𝑦) where 𝑥 = 𝜙(𝑠, 𝑡) and 𝑦 = 𝜓(𝑠, 𝑡) then 𝑢 is treated as a function of 𝑠 and 𝑡 and
𝝏𝒖 𝝏𝒖 𝝏𝒙 𝝏𝒖 𝝏𝒚
 𝝏𝒔
= 𝝏𝒙 𝝏𝒔 + 𝝏𝒚 𝝏𝒔
𝝏𝒖 𝝏𝒖 𝝏𝒙 𝝏𝒖 𝝏𝒚
 𝝏𝒕
= 𝝏𝒙 𝝏𝒕 + 𝝏𝒚 𝝏𝒕
 If 𝑢 = 𝑓(𝑥, 𝑦, 𝑧) where 𝑥 = 𝜙(𝑟, 𝑠, 𝑡), 𝑦 = 𝜓(𝑟, 𝑠, 𝑡) and 𝑧 = 𝜑(𝑟, 𝑠, 𝑡) then
𝝏𝒖 𝝏𝒖 𝝏𝒙 𝝏𝒖 𝝏𝒚 𝝏𝒖 𝝏𝒛
 𝝏𝒓
= 𝝏𝒙 𝝏𝒓 + 𝝏𝒚 𝝏𝒓 + 𝝏𝒛 𝝏𝒓
𝝏𝒖 𝝏𝒖 𝝏𝒙 𝝏𝒖 𝝏𝒚 𝝏𝒖 𝝏𝒛
 𝝏𝒔
= 𝝏𝒙 𝝏𝒔 + 𝝏𝒚 𝝏𝒔 + 𝝏𝒛 𝝏𝒔
𝝏𝒖 𝝏𝒖 𝝏𝒙 𝝏𝒖 𝝏𝒚 𝝏𝒖 𝝏𝒛
 𝝏𝒕
= 𝝏𝒙 𝝏𝒕 + 𝝏𝒚 𝝏𝒕 + 𝝏𝒛 𝝏𝒕
Differentiation of implicit functions:
If 𝑓 𝑥, 𝑦 = 𝑐, 𝑐 being a constant, be an implicit relation between the variables 𝑥 and 𝑦, then
𝝏𝒇
𝒅𝒇 𝝏𝒇 𝝏𝒇 𝒅𝒚 𝒅𝒚 𝝏𝒙
𝒅𝒙
= 𝝏𝒙 + 𝝏𝒚 𝒅𝒙 = 𝟎 ⇒ 𝒅𝒙 = − 𝝏𝒇
𝝏𝒚

Example 10: If 𝑢 = 𝑓 𝑥 − 𝑦, 𝑦 − 𝑧, 𝑧 − 𝑥 , show that

 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 = 0

Solution: Given
𝑢 = 𝑓(𝑥 − 𝑦, 𝑦 − 𝑧, 𝑧 − 𝑥)
Let 𝑟 = 𝑥 − 𝑦, 𝑠 = 𝑦 − 𝑧, 𝑡 = 𝑧 − 𝑥
Then 𝑢 = 𝑓(𝑟, 𝑠, 𝑡).
𝜕𝑢 𝜕𝑢 𝜕𝑟 𝜕𝑢 𝜕𝑠 𝜕𝑢 𝜕𝑡 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢
∴ 𝜕𝑥 = 𝜕𝑟 𝜕𝑥
+ 𝜕𝑠 𝜕𝑥
+ 𝜕𝑡 𝜕𝑥
= 𝜕𝑟
× 1 + 𝜕𝑠 × 0 + 𝜕𝑡 × −1 = 𝜕𝑟
− 𝜕𝑡
Similarly,
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑦
= 𝜕𝑠
− 𝜕𝑟
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑧
= 𝜕𝑡
− 𝜕𝑠
Therefore,
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢
+ + = − + − + − = 0.
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑟 𝜕𝑡 𝜕𝑠 𝜕𝑟 𝜕𝑡 𝜕𝑠

Example 11: If ∅ 𝑥, 𝑦, 𝑧 = 0, show that

𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑦 𝑧
× 𝜕𝑧 𝑥
× 𝜕𝑥 𝑦
= −1

Solution: Given
∅ 𝑥, 𝑦, 𝑧 = 0 (1)
Differentiating (1) partially with respect to 𝑦, keeping 𝑧 as constant, we get
𝜕∅
𝜕∅ 𝜕∅ 𝜕𝑥 𝜕𝑥 𝜕𝑦
𝜕𝑦
+ 𝜕𝑥 𝜕𝑦 𝑧
=0⇒ 𝜕𝑦 𝑧
=− 𝜕∅
𝜕𝑥
Similarly,
𝜕∅
𝜕𝑦 𝜕𝑧
𝜕𝑧 𝑥
=− 𝜕∅
𝜕𝑦
𝜕∅
𝜕𝑧 𝜕𝑥
𝜕𝑥 𝑦
=− 𝜕∅
𝜕𝑧
Therefore,
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑦 𝑧
× 𝜕𝑧 𝑥
× 𝜕𝑥 𝑦
= −1.

𝑑𝑢
Example 12: If 𝑢 = sin(2𝑥 + 𝑦) where 𝑥 = 𝑒 𝑡 and 𝑦 = 𝑡 2 , compute the total derivative 𝑑𝑡
in terms of 𝑡.

Also check the result by direct substitution.

Solution: We have
𝑢 = sin(2𝑥 + 𝑦) where 𝑥 = 𝑒 𝑡 and 𝑦 = 𝑡 2
Therefore,
𝜕𝑢 𝜕𝑢 𝑑𝑥 𝑑𝑦
= 2 cos(2𝑥 + 𝑦) , = cos(2𝑥 + 𝑦) , = 𝑒𝑡 , = 2𝑡
𝜕𝑥 𝜕𝑦 𝑑𝑡 𝑑𝑡
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝑑𝑢 𝜕𝑢 𝑑𝑥 𝜕𝑢 𝑑𝑦
∴ 𝑑𝑡
= 𝜕𝑥 𝑑𝑡
+ 𝜕𝑦 𝑑𝑡
= 2 cos(2𝑥 + 𝑦) 𝑒 𝑡 + 2𝑡 cos(2𝑥 + 𝑦) = 2 𝑒 𝑡 + 𝑡 cos(2𝑒 𝑡 + 𝑡 2 )
Also 𝑢 = sin(2𝑒 𝑡 + 𝑡 2 )
𝑑𝑢
∴ 𝑑𝑡
= cos(2𝑒 𝑡 + 𝑡 2 ) × 2𝑒 𝑡 + 2𝑡 = 2 𝑒 𝑡 + 𝑡 cos(2𝑒 𝑡 + 𝑡 2 ).
Clearly, both results are same.

Jacobian:
𝜕 (𝑢,𝑣)
 If 𝑢 = 𝑓(𝑥, 𝑦) and 𝑣 = 𝑔(𝑥, 𝑦) then the Jacobian of 𝑢 and 𝑣 with respect 𝑥 and 𝑦 is denoted by 𝜕(𝑥 ,𝑦 )
𝑢,𝑣
or 𝐽 𝑥,𝑦
and defined by
𝜕𝑢 𝜕𝑢
𝜕(𝑢, 𝑣) 𝜕𝑥 𝜕𝑦
=
𝜕(𝑥, 𝑦) 𝜕𝑣 𝜕𝑣
𝜕𝑥 𝜕𝑦
 If 𝑢 = 𝑓 𝑥, 𝑦, 𝑧 , 𝑣 = 𝑔(𝑥, 𝑦, 𝑧) and 𝑤 = 𝑕(𝑥, 𝑦, 𝑧) then the Jacobian of 𝑢, 𝑣 and 𝑤 with respect 𝑥, 𝑦
and 𝑧 is given by
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕(𝑢, 𝑣, 𝑤) 𝜕𝑣 𝜕𝑣 𝜕𝑣
=
𝜕(𝑥, 𝑦, 𝑧) 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑤 𝜕𝑤 𝜕𝑤
𝜕𝑥 𝜕𝑦 𝜕𝑧

Properties:
𝜕 (𝑢,𝑣) 𝜕(𝑥,𝑦 )
 If 𝐽 = 𝜕 (𝑥,𝑦 ) and 𝐽 ′ = 𝜕(𝑢,𝑣) then 𝐽 × 𝐽 ′ = 1

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

𝜕 (𝑢,𝑣) 𝜕 (𝑢,𝑣) 𝜕 (𝑟,𝑠)
𝜕(𝑥 ,𝑦 )
= 𝜕 (𝑟,𝑠)
× 𝜕 (𝑥,𝑦 )
𝜕 (𝑥,𝑦 )
Example 13: In polar coordinates, 𝑥 = 𝑟 cos 𝜃 , 𝑦 = 𝑟 sin 𝜃 show that 𝜕 (𝑟,𝜃 ) = 𝑟.

Solution: We have
𝑥 = 𝑟 cos 𝜃 , 𝑦 = 𝑟 sin 𝜃
Therefore,
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕𝑟
= cos 𝜃 , 𝜕𝜃 = −𝑟 sin 𝜃 , 𝜕𝑟 = sin 𝜃 , 𝜕𝜃 = 𝑟 cos 𝜃
𝜕𝑥 𝜕𝑥
𝜕(𝑥 ,𝑦 ) 𝜕𝑟 𝜕𝜃 cos 𝜃 −𝑟 sin 𝜃
∴ 𝜕 (𝑟 ,𝜃 ) = 𝜕𝑦 𝜕𝑦
= = 𝑟 cos2 𝜃 + sin2 𝜃 = 𝑟
sin 𝜃 𝑟 cos 𝜃
𝜕𝑟 𝜕𝜃
𝜕(𝑢,𝑣,𝑤 )
Example 14: If 𝑢 = 𝑥 + 3𝑦 2 − 𝑧 3 , 𝑣 = 4𝑥 2 𝑦𝑧, 𝑤 = 2𝑧 2 − 𝑥𝑦, compute 𝜕 (𝑥,𝑦 ,𝑧)
at 𝑥 = 1, 𝑦 = −1, 𝑧 = 0.

Solution: Here
𝑢 = 𝑥 + 3𝑦 2 − 𝑧 3 , 𝑣 = 4𝑥 2 𝑦𝑧, 𝑤 = 2𝑧 2 − 𝑥𝑦
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Clearly,
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑤 𝜕𝑤 𝜕𝑤
= 1, = 6𝑦, = −3𝑧 2 ; = 8𝑥𝑦𝑧, = 4𝑥 2 𝑧, = 4𝑥 2 𝑦; = −𝑦, = −𝑥, = 4𝑧.
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧

Therefore, at 𝑥 = 1, 𝑦 = −1, 𝑧 = 0
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑤 𝜕𝑤 𝜕𝑤
𝜕𝑥
= 1, 𝜕𝑦 = −6, 𝜕𝑧 = 0; 𝜕𝑥 = 0, 𝜕𝑦 = 0, 𝜕𝑧 = −4; 𝜕𝑥
= 1, 𝜕𝑦 = −1, 𝜕𝑧 = 0
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑣 𝜕𝑣 𝜕𝑣
1 −6 0
𝜕(𝑢 ,𝑣,𝑤 )
∴ = 𝜕𝑥 𝜕𝑦 𝜕𝑧
= 0 0 −4 = 24 − 4 = 20
𝜕 (𝑥,𝑦 ,𝑧)
𝜕𝑤 𝜕𝑤 𝜕𝑤 1 −1 0
𝜕𝑥 𝜕𝑦 𝜕𝑧

Problem Set-III
𝑑𝑢
1. Find the total derivative 𝑑𝑡
in terms of 𝑡 when
(i) 𝑢 = 𝑥 2 + 𝑦 2 where 𝑥 = 𝑎𝑡 2 and 𝑦 = 2𝑎𝑡
𝑦
(ii) 𝑢 = tan−1 𝑥
where 𝑥 = 𝑒 𝑡 − 𝑒 −𝑡 and 𝑦 = 𝑒 𝑡 + 𝑒 −𝑡
𝑦 −𝑥 𝑧−𝑥
2. If 𝑢 = 𝑓 , , show that
𝑥𝑦 𝑥𝑧
𝜕𝑢 2 𝜕𝑢 𝜕𝑢
𝑥2 +𝑦 + 𝑧2 =0
𝜕𝑥 𝜕𝑦 𝜕𝑧
3. If 𝑢 = 𝑓 𝑎𝑥 − 𝑏𝑦, 𝑏𝑦 − 𝑐𝑧, 𝑐𝑧 − 𝑎𝑥 , show that
1 𝜕𝑢 1 𝜕𝑢 1 𝜕𝑢
+ + =0
𝑎 𝜕𝑥 𝑏 𝜕𝑦 𝑐 𝜕𝑧
𝑦 −𝑧 𝑧−𝑥
4. If 𝑢 = 𝑓 𝑒 ,𝑒 , 𝑒 𝑥 −𝑦 , show that
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑥
+ 𝜕𝑦 + 𝜕𝑧 = 0
𝑥 𝑦 𝑧
5. If 𝑢 = 𝑓 , , , show that
𝑦 𝑧 𝑥
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝑥 +𝑦 +𝑧 =0
𝜕𝑥 𝜕𝑦 𝜕𝑧
6. If ∅ 𝑐𝑥 − 𝑎𝑧, 𝑐𝑥 − 𝑏𝑧 = 0, prove that
𝜕𝑧 𝜕𝑧
𝑎 𝜕𝑥 = 𝑏 𝜕𝑦 = 𝑐
𝑧 𝑦
7. If ∅ 𝑥3 𝑥
, = 0, prove that
𝜕𝑧 𝜕𝑧
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 3𝑧
8. If 𝑧 = 𝑓(𝑥, 𝑦) where 𝑥 = 𝑒 𝑢 + 𝑒 −𝑣 and 𝑦 = 𝑒 −𝑢 − 𝑒 𝑣 , prove that
𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
𝜕𝑢
− 𝜕𝑣 = 𝑥 𝜕𝑥 − 𝑦 𝜕𝑦
9. If 𝑢 = 𝑓 𝑟, 𝑠 where 𝑟 = 𝑥 + 𝑎𝑡 and 𝑠 = 𝑦 + 𝑏𝑡, prove that
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝜕𝑡
= 𝑎 𝜕𝑥 + 𝑏 𝜕𝑦
𝜕 (𝑢,𝑣)
10. If 𝑢 = 𝑥 2 − 𝑦 2 , 𝑣 = 2𝑥𝑦 and 𝑥 = 𝑟 cos 𝜃 , 𝑦 = 𝑟 sin 𝜃, find 𝜕 (𝑟,𝜃 ).
𝜕 (𝑥,𝑦 ,𝑧)
11. If 𝑢 = 𝑥𝑦𝑧, 𝑣 = 𝑥 2 + 𝑦 2 + 𝑧 2 , 𝑤 = 𝑥 + 𝑦 + 𝑧, find 𝜕(𝑢,𝑣,𝑤 ).
𝜕 (𝑢,𝑣,𝑤 )
12. If 𝑢 = 𝑥 2 + 𝑦 2 + 𝑧 2 , 𝑣 = 𝑥𝑦 + 𝑦𝑧 + 𝑧𝑥, 𝑤 = 𝑥 + 𝑦 + 𝑧, find .
𝜕 (𝑥 ,𝑦,𝑧)
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Homogeneous Function and Euler's Theorem:

If a function 𝑢 = 𝑓(𝑥, 𝑦) is homogeneous of degree 𝑛 i.e. if 𝑓 𝑡𝑥, 𝑡𝑦 = 𝑡 𝑛 𝑓(𝑥, 𝑦) then

𝝏𝒖 𝝏𝒖
 𝒙 𝝏𝒙 + 𝒚 𝝏𝒚 = 𝒏𝒖
𝝏𝟐 𝒖 𝝏𝟐 𝒖 𝝏𝟐 𝒖
 𝒙𝟐 𝝏𝒙𝟐 + 𝟐𝒙𝒚 𝝏𝒙𝝏𝒚 + 𝒚𝟐 𝝏𝒚𝟐 = 𝒏 𝒏 − 𝟏 𝒖

𝑥 4 +𝑦 4 𝜕𝑢 𝜕𝑢
Example 15: If 𝑢 = log 𝑥+𝑦
, show that 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 3.

Solution: Given
𝑥 4 +𝑦 4 𝑥 4 +𝑦 4
𝑢 = log 𝑥+𝑦
⇒ 𝑒𝑢 = 𝑥 +𝑦
= 𝑓(𝑥, 𝑦), say.

Now,
𝑡 4 𝑥 4 +𝑡 4 𝑦 4 𝑥 4 +𝑦 4
𝑓 𝑡𝑥, 𝑡𝑦 = = 𝑡3 × = 𝑡 3 𝑓(𝑥, 𝑦)
𝑡𝑥 +𝑡𝑦 𝑥+𝑦

Clearly, 𝑓(𝑥, 𝑦) is a homogeneous function in 𝑥 and 𝑦 of degree 𝑛 = 3. Therefore, by Euler’s theorem

𝜕𝑓 𝜕𝑓
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 3𝑓
𝜕 𝜕
⇒𝑥 𝑒𝑢 + 𝑦 𝑒 𝑢 = 3𝑒 𝑢
𝜕𝑥 𝜕𝑦
𝜕𝑢 𝜕𝑢
⇒ 𝑥𝑒 𝑢 𝜕𝑥
+ 𝑦𝑒 𝑢 𝜕𝑦 = 3𝑒 𝑢
𝜕𝑢 𝜕𝑢
⇒ 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 3 (Proved)
𝑥 3 +𝑦 3
Example 16: If 𝑢 = tan−1 𝑥−𝑦
, show that
𝜕𝑢 𝜕𝑢
(i) 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = sin 2𝑢

𝜕2𝑢 𝜕2𝑢 𝜕2𝑢

(ii) 𝑥2 + 2𝑥𝑦 + 𝑦2 = 2 cos 3𝑢 sin 𝑢
𝜕𝑥2 𝜕𝑥𝜕𝑦 𝜕𝑦2

Solution: We have
𝑥 3 +𝑦 3 𝑥 3 +𝑦 3
𝑢 = tan−1 𝑥−𝑦
⇒ tan 𝑢 = 𝑥 −𝑦
= 𝑓(𝑥, 𝑦), say.

Now,
𝑡 3 𝑥 3 +𝑡 3 𝑦 3 𝑥 3 +𝑦 3
𝑓 𝑡𝑥, 𝑡𝑦 = 𝑡𝑥 −𝑡𝑦
= 𝑡2 × 𝑥−𝑦
= 𝑡 2 𝑓(𝑥, 𝑦)

Clearly, 𝑓(𝑥, 𝑦) is a homogeneous function in 𝑥 and 𝑦 of degree 𝑛 = 2. Therefore, by Euler’s theorem

𝜕𝑓 𝜕𝑓
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 2𝑓
𝜕 𝜕
⇒ 𝑥 𝜕𝑥 tan 𝑢 + 𝑦 𝜕𝑦 tan 𝑢 = 2 tan 𝑢
𝜕𝑢 𝜕𝑢
⇒ 𝑥 sec 2 𝑢 𝜕𝑥 + 𝑦 sec 2 𝑢 𝜕𝑦 = 2 tan 𝑢
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝜕𝑢 𝜕𝑢
⇒ 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 2 tan 𝑢 cos2 𝑢 = 2 sin 𝑢 cos 𝑢 = sin 2𝑢 (1)
Differentiating (1) w.r.t. 𝑥, we get
𝜕𝑢 𝜕2𝑢 𝜕2𝑢 𝜕𝑢
𝜕𝑥
+ 𝑥 𝜕 𝑥 2 + 𝑦 𝜕𝑥𝜕𝑦 = 2 cos 2𝑢 𝜕𝑥 (2)
Differentiating (1) w.r.t. 𝑦, we get
𝜕𝑢 𝜕2𝑢 𝜕2𝑢 𝜕𝑢
𝜕𝑦
+ 𝑦 𝜕𝑦 2 + 𝑥 𝜕𝑥𝜕𝑦 = 2 cos 2𝑢 𝜕𝑦 (3)
Multiplying (2) with 𝑥 and (3) with 𝑦 and then adding, we get
𝜕𝑢 𝜕𝑢 𝜕2𝑢 𝜕2𝑢 𝜕2𝑢 𝜕𝑢 𝜕𝑢
𝑥 +𝑦 + 𝑥2 + 2𝑥𝑦 + 𝑦2 = 2 cos 2𝑢 𝑥 +𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑥2 𝜕𝑥𝜕𝑦 𝜕𝑦2 𝜕𝑥 𝜕𝑦
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
⇒ 𝑥2 𝜕 𝑥 2 + 2𝑥𝑦 𝜕𝑥𝜕𝑦 + 𝑦 2 𝜕𝑦 2 = 2 cos 2𝑢 sin 2𝑢 − sin 2𝑢
= sin 4𝑢 − sin 2𝑢
4𝑢−2𝑢 4𝑢+2𝑢
= 2 sin 2
cos 2
= 2 cos 3𝑢 sin 𝑢

Problem Set-IV

𝑥− 𝑦 𝜕𝑢 𝑦 𝜕𝑢
1. If 𝑢 = sin−1 𝑥+ 𝑦
, show by Euler’s theorem that 𝜕𝑥
= − 𝑥 𝜕𝑦 .
𝑥 +2𝑦 +3𝑧
2. If 𝑢 = sin−1 , by Euler’s theorem prove that
𝑥 8 +𝑦 8 +𝑧 8
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 + 𝑧 𝜕𝑧 + 3 tan 𝑢 = 0.
𝑥 3 +𝑦 3 𝜕𝑢 𝜕𝑢 5
3. If 𝑢 = sin−1 𝑥+ 𝑦
, using Euler’s theorem prove that 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 2 tan 𝑢.

Maxima and Minima and Lagrange’s Multipliers

Unconstrained Optimization: Let 𝑢 = 𝑓(𝑥, 𝑦) be a real valued function of 𝑥 and 𝑦 . Let it be

differentiable ∀𝑥 and ∀𝑦.
Procedure:
𝜕𝑢 𝜕𝑢 𝝏𝒖 𝝏𝒖
(i) compute 𝜕𝑥 and 𝜕𝑦 and solve the equations 𝝏𝒙 = 𝟎 and 𝝏𝒚 = 𝟎 to obtain stationary points.

𝜕2𝑢 𝜕 2𝑢 𝜕2𝑢
(ii) compute 𝐴 = 𝜕 𝑥 2 , 𝐵 = 𝜕𝑥𝜕𝑦 and 𝐶 = 𝜕𝑦 2 at each stationary point.
𝜕2𝑢 𝜕2𝑢
𝜕𝑥2 𝜕𝑥𝜕𝑦
(iii) compute 𝑯 𝒙, 𝒚 = 𝑨𝑪 − 𝑩𝟐 =
𝜕2𝑢 𝜕2𝑢
𝜕𝑥𝜕𝑦 𝜕𝑦2

Decision: If at any stationary point

 𝑯 > 0 and 𝑨 > 0 then 𝑓(𝑥, 𝑦) has a minimum value.
 𝑯 > 0 and 𝑨 < 0 then 𝑓(𝑥, 𝑦) has a maximum value.
 𝑯 < 0 then the point is called a saddle point or critical point or point of inflexion.
 𝑯 = 𝟎, no conclusion can be drawn and further investigation is required.
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Example 17: Examine the following function for extreme values:

𝑓 𝑥, 𝑦 = 𝑥 4 + 𝑦 4 − 2𝑥 2 + 4𝑥𝑦 − 2𝑦 2
Solution: We have
𝑓 𝑥, 𝑦 = 𝑥 4 + 𝑦 4 − 2𝑥 2 + 4𝑥𝑦 − 2𝑦 2
Therefore,
𝜕𝑓 𝜕𝑓
𝜕𝑥
= 4𝑥 3 − 4𝑥 + 4𝑦 and 𝜕𝑦 = 4𝑦 3 + 4𝑥 − 4𝑦

𝜕2𝑓 𝜕2𝑓 𝜕2𝑓

𝜕𝑥 2
= 12𝑥 2 − 4, 𝜕𝑥𝜕𝑦 = 4, 𝜕𝑦 2 = 12𝑦 2 − 4

Now,
𝜕𝑓
𝜕𝑥
= 0 ⇒ 4𝑥 3 − 4𝑥 + 4𝑦 = 0 ⇒ 𝑥 3 − 𝑥 + 𝑦 = 0 (1)
𝜕𝑓
= 0 ⇒ 4𝑦 3 + 4𝑥 − 4𝑦 = 0 ⇒ 𝑦 3 + 𝑥 − 𝑦 = 0 (2)
𝜕𝑦

Adding (1) and (2), we get

𝑥 3 + 𝑦 3 = 0 which gives 𝑦 = −𝑥.
Putting 𝑦 = −𝑥 in (1), we get 𝑥 = 0, 2, − 2 and hence 𝑦 = 0, − 2, 2.
Thus the stationary points are 𝑃 0,0 , 𝑄 2, − 2 and 𝑅 − 2, 2 .
Now,
𝜕2𝑓 𝜕2𝑓
𝜕𝑥 2 𝜕𝑥𝜕𝑦 12𝑥 2 − 4 4
𝐻 𝑥, 𝑦 = = = 144𝑥 2 𝑦 2 − 48 𝑥 2 + 𝑦 2
𝜕2𝑓 𝜕2𝑓 4 12𝑦 2 − 4
𝜕𝑥𝜕𝑦 𝜕𝑦 2

(i) At 𝑃 0,0 , 𝐻 𝑥, 𝑦 = 0 ⇒ further investigation is required.

𝜕2𝑓
(ii) At 𝑄 2, − 2 , 𝐻 𝑥, 𝑦 = 384 > 0 and 𝜕𝑥2
= 20 > 0 ⇒ 𝑓(𝑥, 𝑦) has minimum value at

𝑄 2, − 2 and 𝑓𝑚𝑖𝑛 = 0.
𝜕2𝑓
(iii) At 𝑅 − 2, 2 , 𝐻 𝑥, 𝑦 = 384 > 0 and = 20 > 0 ⇒ 𝑓(𝑥, 𝑦) has minimum value at
𝜕𝑥2

𝑅 − 2, 2 and 𝑓𝑚𝑖𝑛 = 0.
Example 18: Examine the following function for extreme values:
𝑓 𝑥, 𝑦 = 𝑥 3 + 𝑦 3 − 3𝑎𝑥𝑦, 𝑎 being a real contant.
Solution: We have
𝑓 𝑥, 𝑦 = 𝑥 3 + 𝑦 3 − 3𝑎𝑥𝑦
Therefore,
𝜕𝑓 𝜕𝑓
𝜕𝑥
= 3𝑥 2 − 3𝑎𝑦 and 𝜕𝑦 = 3𝑦 2 − 3𝑎𝑥
𝜕2𝑓 𝜕2𝑓 𝜕2𝑓
𝜕𝑥 2
= 6𝑥, 𝜕𝑥𝜕𝑦 = −3𝑎, 𝜕𝑦 2 = 6𝑦
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Now,
𝜕𝑓
= 0 ⇒ 3𝑥 2 − 3𝑎𝑦 = 0 ⇒ 𝑥 2 − 𝑎𝑦 = 0 (1)
𝜕𝑥
𝜕𝑓
𝜕𝑦
= 0 ⇒ 3𝑦 2 − 3𝑎𝑥 = 0 ⇒ 𝑦 2 − 𝑎𝑥 = 0 (2)

Subtracting (2) from (1), we get

𝑥 − 𝑦 (𝑥 + 𝑦 − 𝑎) = 0 which gives 𝑦 = 𝑥.
Putting 𝑦 = 𝑥 in (1), we get 𝑥 = 0, 𝑎 and hence 𝑦 = 0, 𝑎.
Thus the stationary points are 𝑃 0,0 and 𝑄 𝑎, 𝑎 .
Now,
𝜕2𝑓 𝜕2𝑓
𝜕𝑥 2 𝜕𝑥𝜕𝑦 6𝑥 −3𝑎
𝐻 𝑥, 𝑦 = = = 36𝑥𝑦 − 9𝑎2
𝜕2𝑓 𝜕2𝑓 −3𝑎 6𝑦
𝜕𝑥𝜕𝑦 𝜕𝑦 2

(i) At 𝑃 0,0 , 𝐻 𝑥, 𝑦 = −9𝑎2 < 0 ⇒ 𝑃 0,0 is a saddle point.

𝜕2𝑓
(ii) At 𝑄 𝑎, 𝑎 , 𝐻 𝑥, 𝑦 = 27𝑎2 > 0 and 𝜕𝑥 2 = 6𝑎 ⇒

 𝑓(𝑥, 𝑦) has minimum value at 𝑄 𝑎, 𝑎 if 𝑎 > 0 and 𝑓𝑚𝑖𝑛 = −𝑎3 .

 𝑓(𝑥, 𝑦) has maximum value at 𝑄 𝑎, 𝑎 if 𝑎 < 0 and 𝑓𝑚𝑎𝑥 = −𝑎3 .
Example 19: Examine the following function for extreme values:
𝑓 𝑥, 𝑦 = 𝑥 3 + 3𝑥𝑦 2 − 15𝑥 2 − 15𝑦 2 + 72𝑥.
Solution: We have
𝑓 𝑥, 𝑦 = 𝑥 3 + 3𝑥𝑦 2 − 15𝑥 2 − 15𝑦 2 + 72𝑥
Therefore,
𝜕𝑓 𝜕𝑓
𝜕𝑥
= 3𝑥 2 + 3𝑦 2 − 30𝑥 + 72 and 𝜕𝑦 = 6𝑥𝑦 − 30𝑦
𝜕2𝑓 𝜕2𝑓 𝜕2𝑓
𝜕𝑥 2
= 6𝑥 − 30, 𝜕𝑥𝜕𝑦 = 6𝑦, 𝜕𝑦 2 = 6𝑥 − 30

Now,
𝜕𝑓
𝜕𝑥
= 0 ⇒ 3𝑥 2 + 3𝑦 2 − 30𝑥 + 72 = 0 ⇒ 𝑥 2 + 𝑦 2 − 10𝑥 + 24 = 0 (1)
𝜕𝑓
= 0 ⇒ 6𝑥𝑦 − 30𝑦 = 0 ⇒ 𝑥𝑦 − 5𝑦 = 0 ⇒ 𝑥 = 5, 𝑦 = 0 (2)
𝜕𝑦

Putting 𝑦 = 0 in (1), we get 𝑥 = 4,6 and 𝑥 = 5 gives from (1), 𝑦 = ±1.

Thus the stationary points are 𝐴 4,0 , 𝐵 6,0 , 𝐶(5,1) and 𝐷 5, −1 .
Now,
𝜕2𝑓 𝜕2𝑓
𝜕𝑥 2 𝜕𝑥𝜕𝑦 6𝑥 − 30 6𝑦
𝐻 𝑥, 𝑦 = = = 36 (𝑥 − 5)2 − 𝑦 2
𝜕2𝑓 𝜕2𝑓 6𝑦 6𝑥 − 30
𝜕𝑥𝜕𝑦 𝜕𝑦 2
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

𝜕2𝑓
(i) At 𝐴 4,0 , 𝐻 𝑥, 𝑦 = 36 > 0 and 𝜕𝑥 2
= −6 < 0 ⇒ 𝑓(𝑥, 𝑦) has maximum value at 𝐴 4,0

and 𝑓𝑚𝑎𝑥 = 112.

𝜕2𝑓
(ii) At 𝐵 6,0 , 𝐻 𝑥, 𝑦 = 36 > 0 and 𝜕 𝑥 2 = 6 > 0 ⇒ 𝑓(𝑥, 𝑦) has minimum value at 𝐵 6,0 and

𝑓𝑚𝑖𝑛 = 108.
(iii) At 𝐶 5,1 , 𝐻 𝑥, 𝑦 = −36 < 0 ⇒ 𝐶 5,1 is a saddle point.
(iv) At 𝐷 5,1 , 𝐻 𝑥, 𝑦 = −36 < 0 ⇒ 𝐶 5, −1 is a saddle point.

Constrained Optimization: Lagrange Multipliers

To find the extreme values of 𝑓 𝑥1 , 𝑥2 , … , 𝑥𝑛 under the constraints
∅1 𝑥1 , 𝑥2 , … , 𝑥𝑛
∅2 𝑥1 , 𝑥2 , … , 𝑥𝑛
………………..
………………..
∅𝑘 𝑥1 , 𝑥2 , … , 𝑥𝑛
we form the equation
𝑔 𝑥1 , 𝑥2 , … , 𝑥𝑛 = 𝑓 𝑥1 , 𝑥2 , … , 𝑥𝑛 + 𝜆1 ∅1 + 𝜆2 ∅2 + ⋯ + 𝜆𝑘 ∅𝑘
The parameters 𝝀𝟏 , 𝝀𝟐 , … , 𝝀𝒌 which are independent of 𝑥1 , 𝑥2 , … , 𝑥𝑛 are called Lagrange multipliers.
We then find the equations:
𝜕𝑔 𝜕𝑔 𝜕𝑔 𝜕𝑔 𝜕𝑔 𝜕𝑔
𝜕𝑥 1
= 0, 𝜕𝑥 = 0, … , 𝜕𝑥 = 0 and 𝜕 𝜆 = 0, 𝜕𝜆 = 0, … , 𝜕𝜆 = 0
2 𝑛 1 2 𝑘
These equations give the necessary conditions for a relative maximum or minimum.

Example 20: Find the dimensions of the rectangular box, open at the top with maximum volume and
given surface area A.
Solution: Let 𝑥, 𝑦, 𝑧 denote the dimensions of the box with 𝑧 as the height. Then
Volume of the box = 𝑥𝑦𝑧 and surface area, 𝐴 = 𝑥𝑦 + 2𝑥𝑧 + 2𝑦𝑧
Hence the problem is
Maximize 𝑓 𝑥, 𝑦, 𝑧 = 𝑥𝑦𝑧 (1)
Subject to
∅ 𝑥, 𝑦, 𝑧 = 𝑥𝑦 + 2𝑥𝑧 + 2𝑦𝑧 − 𝐴 = 0 (2)
Let 𝑔 𝑥, 𝑦, 𝑧 = 𝑓 𝑥, 𝑦, 𝑧 + 𝜆∅ 𝑥, 𝑦, 𝑧 = 𝑥𝑦𝑧 + 𝜆(𝑥𝑦 + 2𝑥𝑧 + 2𝑦𝑧 − 𝐴)
𝜕𝑔
∴ = 0 ⇒ 𝑦𝑧 + 𝜆 𝑦 + 2𝑧 = 0 (3)
𝜕𝑥
𝜕𝑔
𝜕𝑦
= 0 ⇒ 𝑥𝑧 + 𝜆 𝑥 + 2𝑧 = 0 (4)
𝜕𝑔
𝜕𝑧
= 0 ⇒ 𝑥𝑦 + 𝜆 2𝑥 + 2𝑦 = 0 (5)

Multiplying (3) by 𝑥, (4) by 𝑦 and (5) by 𝑧 we get

𝑥𝑦𝑧 + 𝜆 𝑥𝑦 + 2𝑥𝑧 = 𝑥𝑦𝑧 + 𝜆 𝑥𝑦 + 2𝑦𝑧 = 𝑥𝑦𝑧 + 𝜆 2𝑥𝑧 + 2𝑦𝑧
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

⇒ 𝜆 𝑥𝑦 + 2𝑥𝑧 = 𝜆 𝑥𝑦 + 2𝑦𝑧 = 𝜆 2𝑥𝑧 + 2𝑦𝑧

⇒ 𝑥𝑦 + 2𝑥𝑧 = 𝑥𝑦 + 2𝑦𝑧 = 2𝑥𝑧 + 2𝑦𝑧 which give 𝑥 = 𝑦 = 2𝑧 (6)
1 𝐴
∴ from (2) we get 𝑧 = 2 3

𝐴
∴ (6) gives 𝑥 = 𝑦 = 3
1 𝐴 3/2
Therefore, the maximum volume = 𝑥𝑦𝑧 = 2 3

Problem Set-V

1. Find the maximum and minimum values if exist for each of the functions:
(i) 𝑓 𝑥, 𝑦 = 𝑥 3 + 𝑦 3 − 24𝑥𝑦
(ii) 𝑓 𝑥, 𝑦 = 2 𝑥 2 − 𝑦 2 − 𝑥 4 + 𝑦 4
2. Using Lagrange multiplier solve the following problem
𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 + 10
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜
𝑥 + 2𝑦 2 = 3
3. Using Lagrange multiplier solve the following problem
𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 2 + 𝑧 2
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜
𝑥 + 𝑦 + 3𝑧 = 2
5𝑥 + 2𝑦 + 𝑧 = 5
4. Using Lagrange multiplier find a point on the plane 2𝑥 + 3𝑦 − 𝑧 = 5 which is nearest to the origin.
5. Find the dimensions of a rectangular box, open at top, of maximum capacity whose surface area is
432 sq. m.
6. If the sum of three numbers is constant, prove that their product will be maximum when they are
equal.
7. If the perimeter of a triangle is constant, prove that its area will be maximum when it is equilateral.
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Solved GATE Problems

𝜕2𝑓
1. If 𝑓 = 𝑦 𝑥 , what is 𝜕𝑥𝜕𝑦 at 𝑥 = 2, 𝑦 = 1? [GATE 2007]
1
(a) 0 (b) ln 2 (c) 1 (d)
ln 2

Solution:
𝜕𝑓 𝜕2𝑓 𝜕2𝑓
𝑓 = 𝑦 𝑥 ⇒ 𝜕𝑦 = 𝑥. 𝑦 𝑥−1 ⇒ 𝜕𝑥𝜕𝑦 = 𝑦 𝑥 −1 + 𝑥. 𝑦 𝑥−1 . ln 𝑦 ⇒ 𝜕𝑥𝜕𝑦 =1
(2,1)

Ans. (c).
2. If 𝑧 = 𝑓(𝑥, 𝑦), 𝑑𝑧 is equal to [GATE 2000]
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
(a) 𝜕𝑥
𝑑𝑥 + 𝜕𝑦 𝑑𝑦 (b) 𝜕𝑦
𝑑𝑥 + 𝜕𝑥 𝑑𝑦
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
(c) 𝜕𝑥
𝑑𝑥 − 𝜕𝑦 𝑑𝑦 (d) 𝜕𝑦
𝑑𝑥 − 𝜕𝑥 𝑑𝑦
𝜕𝑓 𝜕𝑓
Solution: Chain rule gives 𝑑𝑧 = 𝑑𝑥 + 𝑑𝑦
𝜕𝑥 𝜕𝑦

Ans. (a).
3. If 𝑓(𝑥, 𝑦) = 𝑎0 𝑥 𝑛 + 𝑎1 𝑥 𝑛−1 𝑦 + ⋯ + 𝑎𝑛−1 𝑥𝑦 𝑛−1 + 𝑎𝑛 𝑦 𝑛 where 𝑎𝑖 (𝑖 = 0 to 𝑛) are constants, then
𝜕𝑓 𝜕𝑓
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 is [GATE 2005]
𝑓 𝑛
(a) (b) (c) 𝑛𝑓 (d) 𝑛 𝑓
𝑛 𝑓

Solution:
𝑛 𝑛 −1 𝑛−1 𝑛
𝑓 𝑡𝑥, 𝑡𝑦 = 𝑎0 𝑡𝑥 + 𝑎1 𝑡𝑥 𝑡𝑦 + ⋯ + 𝑎𝑛−1 𝑡𝑥 𝑡𝑦 + 𝑎𝑛 𝑡𝑦 = 𝑡 𝑛 𝑓(𝑥, 𝑦)
Thus 𝑓(𝑥, 𝑦) is a homogeneous function of degree 𝑛 and hence by Euler's theorem we get
𝜕𝑓 𝜕𝑓
𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 = 𝑛𝑓

Ans. (c).

4. The function 𝑓 𝑥, 𝑦 = 2𝑥 2 + 2𝑥𝑦 − 𝑦 3 has [GATE 2002]

(a) only one stationary point at (0,0)
(b) two stationary points at (0,0) and (1/6, −1/3)
(c) two stationary points at (0,0) and (1,−1)
(d) no stationary point

Solution:
𝜕𝑓 𝜕𝑓
𝑓 𝑥, 𝑦 = 2𝑥 2 + 2𝑥𝑦 − 𝑦 3 ⇒ = 4𝑥 + 2𝑦 & = 2𝑥 − 3𝑦 2
𝜕𝑥 𝜕𝑦
𝜕𝑓 𝜕𝑓
∴ 𝜕𝑥
= 0 and 𝜕𝑦 = 0 give 𝑥 = 0, 𝑦 = 0 and 𝑥 = 1/6, 𝑦 = −1/3.
 Multivariate Calculus: MTH-174 Prof. U. Sarkar

Ans. (b).

5. Given a function 𝑓 𝑥, 𝑦 = 4𝑥 2 + 6𝑦 2 − 8𝑥 − 4𝑦 + 8. The optimal value of 𝑓 𝑥, 𝑦 [GATE 2010]

10
(a) is a minimum equal to 3
10
(b) is a maximum equal to 3
8
(c) is a minimum equal to 3
8
(d) is a maximum equal to
3
Solution:
𝜕𝑓 𝜕𝑓
𝑓 𝑥, 𝑦 = 4𝑥 2 + 6𝑦 2 − 8𝑥 − 4𝑦 + 8 ⇒ = 8𝑥 − 8 & = 12𝑦 − 4
𝜕𝑥 𝜕𝑦
𝜕𝑓 𝜕𝑓
∴ 𝜕𝑥
= 0 and 𝜕𝑦 = 0 give 𝑥 = 1, 𝑦 = 1/3.
𝜕2𝑓 𝜕2𝑓 𝜕2𝑢
𝐴 = 𝜕𝑥 2 = 8, 𝐵 = 𝜕𝑥𝜕𝑦 = 0, 𝐶 = 𝜕 𝑦 2 = 12

∴ at (1, 1/3), 𝐴𝐶 − 𝐵2 > 0 and 𝐴 > 0 ⇒ 𝑓 𝑥, 𝑦 has minimum which is

𝑓 1,1/3 = 4 + 6/9 − 8 − 4/3 + 8 = 10/3
Ans. (a).