Department of Mathematics

Multivariable Functions and Partial Differentiation

Topic Learning Objectives:

To recognize the geometrical and physical quantities which depend

on more than one independent variable, evaluate the possible partial
changes due to change in any of these variables or total change due to
changes in all the variables. Study of applications of these concepts
in finding extreme points on a surface and Jacobian as a
transformation factor between different coordinate systems.

Introduction:

The physical quantities like displacement, density, temperature of

metal plates etc. depend on more than one variable. These quantities
vary with space and time, hence rate of change of these lead to
concepts of partial derivatives.

For example, the area of a rectangle of length ‘𝑙’ and breadth ‘𝑏’ is
given by 𝐴 = 𝑙𝑏. This can be represented as a general function 𝑢 =
𝑓 (𝑙, 𝑏). There are two types of changes possible with such a function
resulting in two first order partial changes and one total differential.

The partial change in area of rectangle is due to changes in any one

of the variables either length or breadth. When there are changes in
both the variables at a time, it results in total change in area.
Partial change in area of rectangle due to change in only length
maintaining breadth at the same value:

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

1
 Department of Mathematics

Partial change in area of rectangle due to change in only breadth

maintaining length at same value:

Total change in area of rectangle due to changes in both length and

breadth:

These concepts can be formally expressed as definitions as follows.

Definition: Let 𝑧 = 𝑓(𝑥, 𝑦) be a function of two variables x and y.

𝜕𝑧
The first order partial derivative of z with respect to x, denoted by 𝜕𝑥
𝜕𝑓 𝜕𝑧 𝑓(𝑥+𝛿𝑥,𝑦)−𝑓(𝑥,𝑦)
or 𝜕𝑥or 𝑧𝑥 or 𝑓𝑥 or 𝒑 is defined as 𝜕𝑥
= 𝑙𝑖𝑚 𝛿𝑥
provided
𝛿𝑥→0
the limit exists.
𝜕𝑧
 (𝜕𝑥) is the partial derivative of z with respect to x, treating y as
constant.
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
2
 Department of Mathematics

𝜕𝑧
The first order partial derivative of z with respect to y, denoted by 𝜕𝑦or
𝜕𝑓 𝜕𝑧 𝑓(𝑥,𝑦+𝛿𝑦)−𝑓(𝑥,𝑦)
or zy or fy or q is defined as𝜕𝑦 = 𝑙𝑖𝑚 provided the
𝜕𝑦 𝛿𝑦→0 𝛿𝑦
limit exists.
𝜕𝑧
 (𝜕𝑦)is the partial derivative of z with respect to y, treating x as
constant.

Each of the first order partial derivatives being functions of (x, y), they
can be further differentiated partially with respect to both x and y
resulting in second order partial derivatives.
𝜕 𝜕𝑧 𝜕2𝑧 𝜕2𝑓
The partial derivatives𝜕𝑥 (𝜕𝑥) = 𝜕𝑥 2 or 𝜕𝑥 2or 𝑧𝑥𝑥 or 𝑓𝑥𝑥 or r
𝜕 𝜕𝑧 𝜕2𝑧 𝜕2 𝑓
( ) = 𝜕𝑦 2 or𝜕𝑦 2 or 𝑧𝑦𝑦 or 𝑓𝑦𝑦 or t
𝜕𝑦 𝜕𝑦
𝜕 𝜕𝑧 𝜕2𝑧
( ) = 𝜕𝑥𝜕𝑦or 𝑧𝑦𝑥 or 𝑓𝑦𝑥 or s
𝜕𝑥 𝜕𝑦
𝜕 𝜕𝑧 𝜕2𝑧
and 𝜕𝑦 (𝜕𝑥) = 𝜕𝑦𝜕𝑥or 𝑧𝑥𝑦 or 𝑓𝑥𝑦 or s
are second order Partial derivatives.
𝜕2𝑧 𝜕2𝑧
In most of the cases, it can be verified that = 𝜕𝑦𝜕𝑥
𝜕𝑥𝜕𝑦
(This is true at every point of continuity in any region of interest)

All the rules of differentiation applicable to functions of a single

independent variable are applicable in partial differentiation also, the
only difference is while differentiating partially with respect to one
independent variable all other independent variables are treated as
constants.

Examples:
𝜕𝑧 𝜕𝑧
1. Evaluate 𝜕𝑥 and 𝜕𝑦, if
(a) 𝑧 = 𝑥 2 𝑦 − 𝑥 𝑠𝑖𝑛 𝑥 𝑦

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

3
 Department of Mathematics

(b)𝑥 + 𝑦 + 𝑧 = 𝑙𝑜𝑔 𝑧
Solution:
(a) Consider 𝑧 = 𝑥 2 𝑦 − 𝑥 𝑠𝑖𝑛 𝑥 𝑦
𝜕𝑧
Differentiating z with respect to x, keeping y as a constant, 𝜕𝑥 =
(2𝑥)𝑦 − {𝑥 𝑐𝑜𝑠(𝑥𝑦) (𝑦) + 𝑠𝑖𝑛(𝑥𝑦) (1)}
i.e., 𝑧𝑥 = 2𝑥𝑦 − 𝑥𝑦 𝑐𝑜𝑠 𝑥 𝑦 − 𝑠𝑖𝑛 𝑥 𝑦 = 𝑥𝑦(2 − 𝑐𝑜𝑠 𝑥 𝑦) − 𝑠𝑖𝑛 𝑥 𝑦
Similarly, differentiating z with respect to y keeping x as a constant,
𝜕𝑧
= 𝑥 2 (1) − {𝑥 𝑐𝑜𝑠 𝑥 𝑦(𝑥) + 𝑠𝑖𝑛(𝑥𝑦) (0)}
𝜕𝑦
i.e., 𝑧𝑦 = 𝑥 2 − 𝑥 2 𝑐𝑜𝑠 𝑥 𝑦 = 𝑥 2 (1 − 𝑐𝑜𝑠 𝑥 𝑦)

(b) Consider 𝑥 + 𝑦 + 𝑧 = 𝑙𝑜𝑔 𝑧

i.e., 𝑙𝑜𝑔 𝑧 − 𝑧 = 𝑥 + 𝑦
(1)
Differentiating (1) partially with respect to x, treating y as constant,
1 𝜕𝑧 𝜕𝑧 𝜕𝑧 1
− = 1 + 0 ⇒ ( ) ( − 1) = 1
𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝑧
𝜕𝑧 1−𝑧 𝜕𝑧 𝑧
i.e., 𝜕𝑥 ( 𝑧 ) = 1 or𝜕𝑥 = (1−𝑧)

Similarly, differentiating (1) partially with respect to y, treating x as

1 𝜕𝑧 𝜕𝑧 𝜕𝑧 1 𝜕𝑧 𝑧
constant, 𝑧 𝜕𝑦 − 𝜕𝑦 = 1 + 0 𝑜𝑟 (𝜕𝑦) (𝑧 − 1) = 1  𝜕𝑦 = (1−𝑧)

𝜕𝑧 𝜕𝑧 𝑧
∴ 𝜕𝑥 = 𝜕𝑦 = (1−𝑧).

𝒓𝟐
𝟏 𝝏 𝝏𝜽 𝝏𝜽
2. If 𝜽 = 𝒕𝒏 𝒆− 𝟒𝒕 find the value of n such that 𝒓𝟐 𝝏𝒓 (𝒓𝟐 𝝏𝒓) = .
𝝏𝒕
𝑟2
Solution: Consider 𝜃 = 𝑡 𝑛 𝑒 − 4𝑡 (1)
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
4
 Department of Mathematics

Differentiating (1) with respect to r, treating t as constant,

 𝑟2
−2𝑟 −𝑟 −𝑟2
= 𝑡 𝑛 𝑒 − 4𝑡 ( 4𝑡 ) = 2 𝑡 𝑛−1 𝑒 4𝑡
r
𝜕𝜃 𝑟3 −𝑟2
∴ 𝑟2 = − 𝑡 𝑛−1 𝑒 4𝑡
𝜕𝑟 2
𝜕 𝜕𝜃 −3𝑟 2 𝑛−1 −𝑟 2⁄ 𝑟3 −𝑟 2⁄ 2𝑟
Now 𝜕𝑟 (𝑟 2 𝜕𝑟 ) = 𝑡 𝑒 4𝑡 + (− ) 𝑡 𝑛−1 𝑒 4𝑡 (− 4𝑡 )
2 2
1 𝜕 𝜕𝜃 3 𝑟2 −𝑟 2
∴ (𝑟 2 )= (− 𝑡 𝑛−1 + 𝑡 𝑛−2 ) 𝑒 ⁄4𝑡 (2)
𝑟 2 𝜕𝑟 𝜕𝑟 2 4
Again, differentiating (1) with respect to t, treating r as constant,
𝑟2 −𝑟 2⁄ −𝑟 2⁄
𝜕𝜃 𝑟2 1
= 𝑛𝑡 𝑛−1 𝑒 − 4𝑡 + 𝑡𝑛 𝑒 4𝑡 (4𝑡 2 ) = (𝑛𝑡 𝑛−1 + 4 𝑟 2 𝑡 𝑛−2 ) 𝑒 4𝑡
𝜕𝑡

(3)
1 𝜕 𝜕𝜃 𝜕𝜃
Given 𝑟 2 𝜕𝑟 (𝑟 2 𝜕𝑟 ) = , equation (1) and (2) yield,
𝜕𝑡
3 𝑟2 −𝑟2 𝑟2 −𝑟 2
(− 𝑡 𝑛−1 + 𝑡 𝑛−2 ) 𝑒 4𝑡 = (𝑛𝑡 𝑛−1 + 𝑡 𝑛−2 ) 𝑒 ⁄4𝑡
2 4 4
3 −𝑟 2⁄ 3 3
i.e.(𝑛 + 2) 𝑡 𝑛−1 𝑒 4𝑡 = 0 ⇒ (𝑛 + 2) = 0 𝑜𝑟 𝑛 = − 2

𝝏𝒛 𝝏𝒛
3. If𝒛 = 𝒆𝒂𝒙+𝒃𝒚 𝒇(𝒂𝒙 − 𝒃𝒚), prove that𝒃 𝝏𝒙 + 𝒂 𝝏𝒚 = 𝟐𝒂𝒃𝒛
Solution: Consider 𝑧 = 𝑒 𝑎𝑥+𝑏𝑦 𝑓(𝑎𝑥 − 𝑏𝑦) (1)
Differentiating (1) with respect to x using product and chain rules,
𝜕𝑧
= {𝑒 𝑎𝑥+𝑏𝑦 𝑓 1 (𝑎𝑥 − 𝑏𝑦)(𝑎)} + {𝑓(𝑎𝑥 − 𝑏𝑦)𝑒 𝑎𝑥+𝑏𝑦 (𝑎)}
𝜕𝑥
𝜕𝑧
= 𝑎𝑒 𝑎𝑥+𝑏𝑦 {𝑓(𝑎𝑥 − 𝑏𝑦) + 𝑓 ′ (𝑎𝑥 − 𝑏𝑦)} (2)
𝜕𝑥
Differentiating (1) with respect to y using product and chain rules,
𝜕𝑧
= {𝑒 𝑎𝑥+𝑏𝑦 𝑓 ′ (𝑎𝑥 − 𝑏𝑦)(−𝑏)} + {𝑓(𝑎𝑥 − 𝑏𝑦)𝑒 𝑎𝑥+𝑏𝑦 (𝑏)}
𝜕𝑦

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

5
 Department of Mathematics

𝜕𝑧
= 𝑏𝑒 𝑎𝑥+𝑏𝑦 {𝑓(𝑎𝑥 − 𝑏𝑦) − 𝑓 ′ (𝑎𝑥 − 𝑏𝑦)} (3)
𝜕𝑦
Multiplying equation (2) by b and equation (3) by a, and adding
𝜕𝑧 𝜕𝑧
𝑏 𝜕𝑥 + 𝑎 𝜕𝑦 = 𝑎𝑏𝑒 𝑎𝑥+𝑏𝑦 [𝑓 ′ (𝑎𝑥 + 𝑏𝑦) + 𝑓(𝑎𝑥 − 𝑏𝑦) + 𝑓(𝑎𝑥 −
𝑏𝑦) − 𝑓 ′ (𝑎𝑥 − 𝑏𝑦)] = 𝑎𝑏𝑒 𝑎𝑥+𝑏𝑦 [2𝑓(𝑎𝑥 − 𝑏𝑦)] =
2𝑎𝑏[𝑒 𝑎𝑥+𝑏𝑦 𝑓(𝑎𝑥 − 𝑏𝑦)] = 2𝑎𝑏𝑧.
4. If 𝒖 = 𝒆𝒓 𝒄𝒐𝒔 𝜽 𝒄𝒐𝒔(𝒓 𝒔𝒊𝒏 𝜽) , 𝒗 = 𝒆𝒓 𝒄𝒐𝒔 𝜽 𝒔𝒊𝒏(𝒓 𝒔𝒊𝒏 𝜽), prove that
𝝏𝒖 𝟏 𝝏𝒗 𝝏𝒗 𝟏 𝝏𝒖
= 𝒓 𝝏𝜽&𝝏𝒓 = − 𝒓 𝝏𝜽
𝝏𝒓
Solution: Consider𝑢 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃), differentiating u with
𝜕𝑢
respect to r and θ partially, 𝜕𝑟 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {− 𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) (𝑠𝑖𝑛 𝜃)} +
𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) {𝑒 𝑟 𝑐𝑜𝑠 𝜃 (𝑐𝑜𝑠 𝜃)}
𝜕𝑢
i.e.,𝜕𝑟 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) 𝑐𝑜𝑠 𝜃 − 𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) 𝑠𝑖𝑛 𝜃}
𝜕𝑢
= 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃 + 𝜃)}
𝜕𝑟
(1)
𝜕𝑢
and𝜕𝜃 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {− 𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) (𝑟 𝑐𝑜𝑠 𝜃)} +
𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) {𝑒 𝑟 𝑐𝑜𝑠 𝜃 (−𝑟 𝑠𝑖𝑛 𝜃)}
= −𝑟𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) 𝑐𝑜𝑠 𝜃 + 𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) 𝑠𝑖𝑛 𝜃}
1 𝜕𝑢
i.e., − 𝑟 𝜕𝜃 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃 + 𝜃)} (2)
Again consider𝑣 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 𝑠𝑖𝑛( 𝑟 𝑠𝑖𝑛 𝜃)
Differentiating v with respect to r & θ partially,
𝜕𝑣
= 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) 𝑠𝑖𝑛 𝜃} + 𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) {𝑒 𝑟 𝑐𝑜𝑠 𝜃 (𝑐𝑜𝑠 𝜃)}
𝜕𝑟
= 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) 𝑐𝑜𝑠 𝜃 + 𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) 𝑠𝑖𝑛 𝜃}

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

6
 Department of Mathematics

𝜕𝑣
i.e., 𝜕𝑟 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃 + 𝜃)}
𝜕𝑣
(3)𝜕𝜃 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) (𝑟 𝑐𝑜𝑠 𝜃)} +
𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) {𝑒 𝑟 𝑐𝑜𝑠 𝜃 (−𝑟 𝑠𝑖𝑛 𝜃)}
= 𝑟𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃) 𝑐𝑜𝑠 𝜃 − 𝑠𝑖𝑛(𝑟 𝑠𝑖𝑛 𝜃) 𝑠𝑖𝑛 𝜃}
1 𝜕𝑣
i.e., 𝑟 𝜕𝜃 = 𝑒 𝑟 𝑐𝑜𝑠 𝜃 {𝑐𝑜𝑠(𝑟 𝑠𝑖𝑛 𝜃 + 𝜃)} (4)
𝜕𝑢 1 𝜕𝑣
Thus, from equations (1) and (4), = , and from equations (2)
𝜕𝑟 𝑟 𝜕𝜃
𝜕𝑣 1 𝜕𝑢
and (3) 𝜕𝑟 = − 𝑟 𝜕𝜃.

Examples:
𝝏𝟐 𝒖 𝝏𝟐 𝒖
For the following functions, verify that𝝏𝒙𝝏𝒚 = 𝝏𝒚𝝏𝒙, where𝒖 =
𝒇(𝒙, 𝒚).
1.𝑢 = 𝑥 𝑦
𝑦
2.𝑢 = 𝑠𝑖𝑛−1 ( ⁄𝑥)
3.𝑢 = 𝑒 𝑥 (𝑥 𝑠𝑖𝑛 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦)

Solution: 1. Differentiating𝑢 = 𝑥 𝑦 partially with respect to y,

𝜕𝑢 𝑑
= 𝑥 𝑦 𝑙𝑜𝑔 𝑥 (𝑠𝑖𝑛 𝑐 𝑒 (𝑎 𝑥 ) = 𝑎 𝑥 𝑙𝑜𝑔 𝑎)
𝜕𝑦 𝑑𝑥
Differentiating the above equation partially with respect to x,
𝜕 𝜕𝑢 𝜕
( )= (𝑥 𝑦 𝑙𝑜𝑔 𝑥)
𝜕𝑥 𝜕𝑦 𝜕𝑥
𝜕2𝑢 1 𝜕2𝑢
i.e., 𝜕𝑥𝜕𝑦 = 𝑥 𝑦 (𝑥) + 𝑙𝑜𝑔 𝑥 (𝑦𝑥 𝑦−1 ) or = 𝑥 𝑦−1 (1 + 𝑦 𝑙𝑜𝑔 𝑥)
𝜕𝑥𝜕𝑦
(1)
𝜕𝑢
Again, differentiating𝑢 = 𝑥 𝑦 partially with respect to x, = 𝑦𝑥 𝑦−1
𝜕𝑥
so that
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
7
 Department of Mathematics

𝜕 𝜕𝑢 𝜕
( )= (𝑦𝑥 𝑦−1 )
𝜕𝑦 𝜕𝑥 𝜕𝑦
𝜕2𝑢
i.e., 𝜕𝑦𝜕𝑥 = 𝑦(𝑥 𝑦−1 𝑙𝑜𝑔 𝑥) + 𝑥 𝑦−1 (1)or
(2)
𝜕2𝑢 𝜕2𝑢
From (1) and (2),𝜕𝑥𝜕𝑦 = 𝜕𝑦𝜕𝑥
𝑦
Solution: 2. Differentiating𝑢 = 𝑠𝑖𝑛−1 ( ⁄𝑥) partially with respect to
x,
𝜕𝑢 1 −𝑦 −𝑦
= ( ⁄𝑥 2 ) =
𝜕𝑥 2 𝑥√𝑥 2 − 𝑦 2
√ 1 − (𝑦 )
𝑥

𝑥√𝑥 2 − 𝑦 2 (1) − 𝑦 {𝑥 (1⁄ ) (−2𝑦)}

𝜕 𝜕𝑢 2√𝑥 2 − 𝑦 2
∴ ( )=−
𝜕𝑦 𝜕𝑥 𝑥 2 (𝑥 2 − 𝑦 2 )
𝜕2𝑢 𝑥(𝑥 2 −𝑦 2 )+𝑥𝑦 2 𝑥
i.e., = −{ 3 }=− 3
𝜕𝑦𝜕𝑥 𝑥 2 (𝑥 2 −𝑦 2 ) ⁄2 (𝑥 2 −𝑦 2 ) ⁄2
(1)
𝑦⁄ −1
Again, differentiating𝑢 = 𝑠𝑖𝑛 ( 𝑥) partially with respect to y,
𝜕𝑢 1 1 1
= 2 ( ⁄𝑥 ) = 2 2
𝜕𝑦 √𝑥 −𝑦
√1−(𝑦⁄𝑥)
1
√𝑥 2 −𝑦 2 (0)−(1) (2𝑥)
𝜕 𝜕𝑢 𝜕2𝑢 2√𝑥2−𝑦2 𝑥
∴ 𝜕𝑥 (𝜕𝑦) = 𝜕𝑥𝜕𝑦 = − =− 3 (2)
𝑥 2 −𝑦 2 (𝑥 2 −𝑦 2 )2
𝜕2𝑢 𝜕2𝑢
From (1) and (2), = 𝜕𝑦𝜕𝑥
𝜕𝑥𝜕𝑦
Solution: 3. Differentiating 𝑢 = 𝑒 𝑥 (𝑥 𝑠𝑖𝑛 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦)with respect
𝜕𝑢
to y, 𝜕𝑦 = 𝑒 𝑥 (𝑥 𝑐𝑜𝑠 𝑦 − 𝑦 𝑐𝑜𝑠 𝑦 − 𝑠𝑖𝑛 𝑦)

Differentiating with respect to x,

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

8
 Department of Mathematics

𝜕 𝜕𝑢
( ) = 𝑒 𝑥 (𝑥 𝑐𝑜𝑠 𝑦 − 𝑦 𝑐𝑜𝑠 𝑦 − 𝑠𝑖𝑛 𝑦) + 𝑒 𝑥 𝑐𝑜𝑠 𝑦
𝜕𝑥 𝜕𝑦
𝜕2𝑢
or 𝜕𝑥𝜕𝑦 = 𝑒 𝑥 {(1 + 𝑥) 𝑐𝑜𝑠 𝑦 − 𝑦 𝑐𝑜𝑠 𝑦 − 𝑠𝑖𝑛 𝑦} (1)
differentiating 𝑢 with respect to 𝑥,
𝜕𝑢
= 𝑒 𝑥 {𝑥 𝑠𝑖𝑛 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦} + 𝑒 𝑥 𝑠𝑖𝑛 𝑦
𝜕𝑥
𝜕𝑢
= 𝑒 𝑥 {(1 + 𝑥) 𝑠𝑖𝑛 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦}
𝜕𝑥
Differentiating above expression with respect to y,
𝜕 𝜕𝑢 𝜕2𝑢
( ) = 𝜕𝑦𝜕𝑥 = 𝑒 𝑥 {(1 + 𝑥) 𝑐𝑜𝑠 𝑦 − 𝑦 𝑐𝑜𝑠 𝑦 − 𝑠𝑖𝑛 𝑦}
𝜕𝑦 𝜕𝑥
(2)
𝜕2𝑢 𝜕2𝑢
From (1) and (2), 𝜕𝑥𝜕𝑦 = 𝜕𝑦𝜕𝑥
Examples:
2𝑥𝑦 𝜕2𝑢 𝜕2 𝑢
1. If𝑢 = 𝑡𝑎𝑛−1 (𝑥 2 −𝑦 2), prove that 𝜕𝑥 2 + 𝜕𝑦 2 = 0

2𝑥𝑦
Solution: Differentiating𝑢 = 𝑡𝑎𝑛−1 (𝑥 2 −𝑦 2) partially with respect to
x,
𝜕𝑢 1 (𝑥 2 − 𝑦 2 )(2𝑦) − 2𝑥𝑦(2𝑥)
= 2{ }
𝜕𝑥 2𝑥𝑦 (𝑥 2 − 𝑦 2 )2
1 + (𝑥 2 −𝑦 2)
𝜕𝑢 −2𝑦
i.e. = (𝑥 2 Differentiating partially this with respect to x,
𝜕𝑥 +𝑦 2 )
𝜕 2𝑢 𝜕 −2𝑦 (𝑥 2 + 𝑦 2 )(0) − (−2𝑦)(2𝑥)
= ( ) =
𝜕𝑥 2 𝜕𝑥 𝑥 2 + 𝑦 2 (𝑥 2 + 𝑦 2 )2

𝜕2𝑢 4𝑥𝑦
i.e., 𝜕𝑥 2 = (𝑥 2+𝑦 2)2 (1)

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

9
 Department of Mathematics

2𝑥𝑦
Again differentiating 𝑢 = 𝑡𝑎𝑛−1 (𝑥 2 −𝑦 2) partially with respect to y,

𝜕𝑢 1 (𝑥 2 − 𝑦 2 )(2𝑥) − 2𝑥𝑦(−2𝑦)
= 2{ }
𝜕𝑦 2𝑥𝑦 (𝑥 2 − 𝑦 2 )2
1 + (𝑥 2 −𝑦 2)
𝜕𝑢 2𝑥
= (𝑥 2+𝑦 2) Differentiating partially this with respect to y,
𝜕𝑦
𝜕2𝑢 𝜕 2𝑥 (𝑥 2 −𝑦 2 )(0)−(2𝑥)(2𝑦)
= ( )=
𝜕𝑥 2 𝜕𝑥 𝑥 2 +𝑦 2 (𝑥 2 +𝑦 2 )2
𝜕2 𝑢 −4𝑥𝑦
i.e., 𝜕𝑦 2 = (𝑥 2 +𝑦 2 )2
(2)
𝜕2𝑢 𝜕2 𝑢
Adding equations (1) and (2) we get + 𝜕𝑦 2 = 0, as desired.
𝜕𝑥 2
Note:
𝜕2𝑢 𝜕2𝑢
(a) The equation𝜕𝑥 2 + 𝜕𝑦 2 = 0 is known as Laplace’s equation in
two dimensions which has variety of applications in field theory.
𝜕2 𝑢 𝜕2𝑢 𝜕2𝑢
(b) The equation + + = 0 is known as Laplace’s
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑧 2
equation in three dimensions where 𝑢 = 𝑢(𝑥, 𝑦, 𝑧).
(c) Any function u(x, y) satisfying Laplace equation is called
harmonic function.
𝜕2𝑢 𝜕2𝑢
2. If 𝑢 = 𝑓(𝑥 + 𝑎𝑦) + 𝑔(𝑥 − 𝑎𝑦), show that = 𝑎2 𝜕𝑥 2 .
𝜕𝑦 2

Solution: Differentiating 𝑢 = 𝑓(𝑥 + 𝑎𝑦) + 𝑔(𝑥 − 𝑎𝑦), partially

with respect to x,
𝜕𝑢
= 𝑓 ′ (𝑥 + 𝑎𝑦)(1) + 𝑔′ (𝑥 − 𝑎𝑦)(1)
𝜕𝑥
Again differentiating partially with respect to x,
𝜕 2𝑢
= 𝑓 ″ (𝑥 + 𝑎𝑦)(1) + 𝑔″ (𝑥 − 𝑎𝑦)(1)
𝜕𝑥 2
𝜕2𝑢
𝑎2 𝜕𝑥 2 = 𝑎2 [𝑓 ″ (𝑥 + 𝑎𝑦) + 𝑔″ (𝑥 − 𝑎𝑦)] (1)
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
10
 Department of Mathematics

Next, differentiating 𝑢 = 𝑓(𝑥 + 𝑎𝑦) + 𝑔(𝑥 − 𝑎𝑦), partially with

respect to y
𝜕𝑢
= 𝑓 ′ (𝑥 + 𝑎𝑦)𝑎 + 𝑔.′ (𝑥 − 𝑎𝑦)(−𝑎)
𝜕𝑦
Again, differentiating partially with respect to y,
𝜕 2𝑢
= 𝑓 ″ (𝑥 + 𝑎𝑦)(𝑎2 ) + 𝑔″ (𝑥 − 𝑎𝑦)(𝑎2 )
𝜕𝑥 2
𝜕2 𝑢
i.e., 𝜕𝑦 2 = 𝑎2 [𝑓 ″ (𝑥 + 𝑎𝑦) + 𝑔″ (𝑥 − 𝑎𝑦)] (2)
𝜕2𝑢 𝜕2𝑢
From equations (1) and (2), 𝜕𝑦 2 = 𝑎2 𝜕𝑥 2
Note:
𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
(a) The equation 𝜕𝑦2 = 𝑎2 𝜕𝑥 2 or the equation 𝜕𝑡 2
= 𝑎2 𝜕𝑥 2 is known
as one-dimensional wave equation.
𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
(b) The equation, 𝜕𝑡 2 = 𝑎2 (𝜕𝑥 2 + 𝜕𝑦2 + 𝜕𝑧2 ) is known as three-
dimensional wave equation
(c) The solution of the type 𝑢 = 𝑓(𝑥 + 𝑎𝑦) + 𝑔(𝑥 − 𝑎𝑦) is called
D’Alembert’s solution of wave equation.

𝜕𝑢 𝜕2𝑢
3. If 𝑢 = 𝑒 −2𝑡 𝑐𝑜𝑠 3 𝑥, find the value of ‘c’ such that = 𝑐 2 𝜕𝑥 2
𝜕𝑡
−2𝑡
Solution: Differentiating𝑢 = 𝑒 𝑐𝑜𝑠 3 𝑥, partially with respect to t
𝜕𝑢 −2𝑡
= −2𝑒 𝑐𝑜𝑠 3 𝑥 (1)
𝜕𝑡
−2𝑡
Next, differentiating 𝑢 = 𝑒 𝑐𝑜𝑠 3 𝑥 partially with respect to x,
𝜕𝑢
= −3𝑒 −2𝑡 𝑠𝑖𝑛 3 𝑥
𝜕𝑥
Again, differentiating partially with respect to x,
𝜕 2𝑢 −2𝑡
2
= −9𝑒 𝑐𝑜𝑠 3 𝑥
𝜕𝑥

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

11
 Department of Mathematics

𝜕𝑢 𝜕2𝑢
∴ = 𝑐 2 𝜕𝑥 2 ⇒ −2𝑒 −2𝑡 𝑐𝑜𝑠 3 𝑥 = 𝑐 2 9𝑒 −2𝑡 𝑐𝑜𝑠 3 𝑥 ⇒ 𝑐 2 =
𝜕𝑡
2 2
(9)or𝑐 = √9
Note:
𝜕𝑢 𝜕2𝑢
The equation 𝜕𝑡 = 𝑐 2 𝜕𝑥 2 is called one-dimensional heat equation
𝜕𝑢 𝜕2𝑢 𝜕2𝑢 𝜕2 𝑢
and the equation = 𝑐 2 (𝜕𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2 ) is known as three-
𝜕𝑡
dimensional heat equation. The constant c in the equation is called
dissipation coefficient.

4. If 𝑢 = 𝑙𝑜𝑔(𝑥 3 + 𝑦 3 + 𝑧 3 − 3𝑥𝑦𝑧), prove that

𝜕𝑢 𝜕𝑢 𝜕𝑢 3
(a) 𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧 = (𝑥+𝑦+𝑧)
𝜕 𝜕 𝜕 2 9
(b) (𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧) 𝑢 = − (𝑥+𝑦+𝑧)2
Solution:
(a) Differentiating u partially with respect to x, y, z
𝜕𝑢 3𝑥 2 −3𝑦𝑧
= 𝑥 3 +𝑦 3+𝑧 3−3𝑥𝑦𝑧 ,
𝜕𝑥
𝜕𝑢 3𝑦 2 −3𝑧𝑥
similarly, 𝜕𝑦 = 𝑥 3 +𝑦 3+𝑧 3−3𝑥𝑦𝑧 (u is symmetric function)
𝜕𝑢 3𝑧 2 −3𝑥𝑦
and 𝜕𝑧 = 𝑥 3+𝑦 3+𝑧 3−3𝑥𝑦𝑧
adding these partial derivatives
𝜕𝑢 𝜕𝑢 𝜕𝑢 3(𝑥 2 + 𝑦 2 + 𝑧 2 − 𝑥𝑦 − 𝑦𝑧 − 𝑧𝑥)
+ + =
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝑥 3 + 𝑦 3 + 𝑧 3 − 3𝑥𝑦𝑧
3(𝑥 2 + 𝑦 2 + 𝑧 2 − 𝑥𝑦 − 𝑦𝑧 − 𝑧𝑥)
=
(𝑥 + 𝑦 + 𝑧)(𝑥 2 + 𝑦 2 + 𝑧 2 − 𝑥𝑦 − 𝑦𝑧 − 𝑧𝑥)
3
=( )
𝑥+𝑦+𝑧

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

12
 Department of Mathematics

(b) By definition, we have

𝜕 𝜕 𝜕 2 𝜕 𝜕 𝜕 𝜕 𝜕 𝜕
( + + ) 𝑢=( + + )( + + )𝑢
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕 𝜕 𝜕 𝜕𝑢 𝜕𝑢 𝜕𝑢
= (𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧) (𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧 )
𝜕 𝜕 𝜕 3
=( + 𝜕𝑦 + 𝜕𝑧) (𝑥+𝑦+𝑧)
𝜕𝑥
𝜕 3 𝜕 3 𝜕 3
= ( )+ ( )+ ( )
𝜕𝑥 𝑥+𝑦+𝑧 𝜕𝑦 𝑥+𝑦+𝑧 𝜕𝑧 𝑥+𝑦+𝑧
−3 −3 −3 9
= ((𝑥+𝑦+𝑧)2)+((𝑥+𝑦+𝑧)2)+((𝑥+𝑦+𝑧)2)=− (𝑥+𝑦+𝑧)2

𝜕2𝑢 𝜕2 𝑢
5. If 𝑢 = 𝑓(𝑟), where𝑟 = √𝑥 2 + 𝑦 2 + 𝑧 2 , show that 𝜕𝑥 2 + 𝜕𝑦 2 +
𝜕2𝑢 2
= 𝑓 11 (𝑟) + 𝑟 𝑓 1 (𝑟)
𝜕𝑧 2
Solution: By data 𝑟 2 = 𝑥 2 + 𝑦 2 + 𝑧 2 differentiating partially with
𝜕𝑟 𝜕𝑟 𝑥
respect to x, 2𝑟 𝜕𝑥 = 2𝑥 or 𝜕𝑥 = (𝑟 ).
𝜕𝑟 𝑦 𝜕𝑟 𝑧
Similarly, we get 𝜕𝑦 = (𝑟 ) and = (𝑟)
𝜕𝑧
Now, differentiating𝑢 = 𝑓(𝑟), partially with respect to x,
𝜕𝑢 𝜕𝑟 𝜕𝑟 𝑥
= 𝑓 1 (𝑟) 𝜕𝑥as𝜕𝑥 = (𝑟 ),
𝜕𝑥
𝜕𝑢 𝑥
= 𝑓 1 (𝑟) ( )
𝜕𝑥 𝑟
𝜕𝑢 𝑓 1 (𝑟)
Differentiating 𝜕𝑥 = 𝑥 { }partially with respect to x,
𝑟
2
𝜕 𝑢 1 11 𝜕𝑟 1
1 𝜕𝑟 𝑓 1 (𝑟)
= 𝑥 { 𝑓 (𝑟) + 𝑓 (𝑟) (− 2 ) } +
𝜕𝑥 2 𝑟 𝜕𝑥 𝑟 𝜕𝑥 𝑟
𝜕2𝑢 1 𝑥 1 𝑥 𝑓 1 (𝑟)
i.e. 𝜕𝑥 2 = 𝑥 {𝑟 𝑓 11 (𝑟) (𝑟 ) + 𝑓 1 (𝑟) (− 𝑟 2 ) (𝑟 )} + 𝑟
𝜕2 𝑢 𝑥 𝑓 11 (𝑟) 𝑓 1 (𝑟) 𝑓 1 (𝑟)
or 𝜕𝑥 2 = 𝑥 (𝑟 ) { − }+
𝑟 𝑟2 𝑟
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
13
 Department of Mathematics

𝜕2𝑢 𝑥2 𝑓 1 (𝑟) 𝑥2
= 𝑓 11 (𝑟) + {1 − } (1)
𝜕𝑥 2 𝑟2 𝑟 𝑟2
𝜕2𝑢 𝑦2 𝑓 1 (𝑟) 𝑦2
= 𝑓 11 (𝑟) + {1 − } (2)
𝜕𝑦 2 𝑟2 𝑟 𝑟2
Similarly,
𝜕2𝑢 𝑧2 𝑓 1 (𝑟) 𝑧2
= 𝑟 2 𝑓 11 (𝑟) + {1 − } (3)
𝜕𝑧 2 𝑟 𝑟2
Adding equations. (1), (2) and (3)

𝜕 2𝑢 𝜕 2𝑢 𝜕 2𝑢
+ +
𝜕𝑥 2 𝜕𝑦 2 𝜕𝑧 2
𝑓 11 (𝑟) 2
= (𝑥 + 𝑦 2 + 𝑧 2 )
𝑟2
𝑓 1 (𝑟) 𝑥2 + 𝑦2 + 𝑧2
+ {3 − }
𝑟 𝑟2

𝑓 1 (𝑟) 𝑓 1 (𝑟)
= 𝑓 11 (𝑟) + (3 − 1) = 𝑓 11 (𝑟) + 2
𝑟 𝑟

Exercise:
𝜕2 𝑧 𝜕2𝑧
1. Verify 𝜕𝑥𝜕𝑦 = 𝜕𝑦𝜕𝑥 where
(i) 𝑧 = 𝑥 3 + 𝑦 3 − 3𝑎𝑥𝑦 (ii)𝑧 = 𝑡𝑎𝑛−1(𝑥 2 + 𝑦 2 )
(iii)𝑧 = 𝑙𝑜𝑔(𝑥 𝑠𝑖𝑛 𝑦 + 𝑦 𝑠𝑖𝑛 𝑥)
𝑦 𝜕2𝑢 𝑥 2 −𝑦 2
2. If𝑢 = 𝑥 2 𝑡𝑎𝑛−1( ⁄𝑥) − 𝑦 2 𝑡𝑎𝑛−1 𝑥⁄𝑦 , prove that𝜕𝑦𝜕𝑥 = 𝑥 2 +𝑦 2
𝑥 2 +𝑦 2 𝜕𝑧 𝜕𝑧 2 𝜕𝑧 𝜕𝑧
3. If𝑧 = 𝑥 , show that (𝜕𝑥 − 𝜕𝑦) = 4 (1 − 𝜕𝑥 − 𝜕𝑦)
+𝑦
𝜕𝑢 𝜕𝑢 𝜕𝑢
4. If 𝑢 = (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥), show that 𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧 = 0

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

14
 Department of Mathematics

𝜕2 𝑧 𝜕2𝑧 1 𝜕𝑧 2 𝜕𝑧 2
5. If 𝑧 = 𝑒 𝑥𝑦 , prove that + 𝜕𝑦 2 = 𝑧 [(𝜕𝑥) + (𝜕𝑦 2) ]
𝜕𝑥 2
𝑦 𝑧 𝜕𝑢 𝜕𝑢 𝜕𝑢
6. If𝑢 = 𝑧 + 𝑥, prove that 𝑥 𝜕𝑥 + 𝑦 𝜕𝑦 + 𝑧 𝜕𝑧 = 0
𝜕2𝑟 𝜕2𝑟 1 𝜕𝑟 2
7. If 𝑥 = 𝑟 𝑐𝑜𝑠 𝜃 , 𝑦 = 𝑟 𝑠𝑖𝑛 𝜃, show that𝜕𝑥 2 + 𝜕𝑦 2 = 𝑟 [(𝜕𝑥) +
𝜕𝑟 2
(𝜕𝑦) ]

Total derivatives, Differentiation of Composite and Implicit

functions

Total differential and Total derivative:

Let 𝑧 = 𝑓(𝑥, 𝑦) be a differentiable function of two variables, x and y

then total differential (or exact differential) ‘dz’ is defined by 𝑑𝑧 =
𝜕𝑓 𝜕𝑓
𝑑𝑥 + 𝑑𝑦. (1)
𝜕𝑥 𝜕𝑦
Further, if 𝑧 = 𝑓(𝑥, 𝑦) where𝑥 = 𝑥(𝑡),𝑦 = 𝑦(𝑡) i.e. x and y are
themselves functions of an independent variable t, then total
derivative of z is given by
dz f dx f dy
= +
dt x dt y dt
(2)

Similarly, the total differential of a function 𝑢 = 𝑓(𝑥, 𝑦, 𝑧) is defined

by
𝜕𝑓 𝜕𝑓
𝑑𝑢 = 𝜕𝑥 𝑑𝑥 + 𝜕𝑦 𝑑𝑦 +
𝜕𝑓
𝑑𝑧
𝜕𝑧
(3)

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

15
 Department of Mathematics

Further, if 𝑢 = 𝑓(𝑥, 𝑦, 𝑧) and if 𝑥 = 𝑥(𝑡),𝑦 = 𝑦(𝑡),𝑧 = 𝑧(𝑡), then the

𝑑𝑢 𝜕𝑓 𝑑𝑥 𝜕𝑓 𝑑𝑦 𝜕𝑓 𝑑𝑧
total derivative of u is given by = 𝜕𝑥 𝑑𝑡 + 𝜕𝑦 𝑑𝑡 + 𝜕𝑧 𝑑𝑡 .
𝑑𝑡
(4)

Differentiation of implicit functions:

An implicit function with x as an independent variable and y as the
dependent variable is generally of the form𝑧 = 𝑓(𝑥, 𝑦) = 0. This
𝑑𝑧 𝑑𝑓 𝑑𝑧
gives (𝑑𝑥) = (𝑑𝑥 ) = 0. Then, by virtue of expression (2) above, 𝑑𝑥 =
𝜕𝑓 𝜕𝑓 𝑑𝑦 𝑑𝑓 𝜕𝑓 𝜕𝑓 𝑑𝑦
+ 𝜕𝑦 𝑑𝑥 or = 𝜕𝑥 + 𝜕𝑦 𝑑𝑥
𝜕𝑥 𝑑𝑥
𝜕𝑓 𝜕𝑓 𝑑𝑦 𝑑𝑦
and hence 0 = 𝜕𝑥 + 𝜕𝑦 𝑑𝑥 , so that 𝑑𝑥 =
𝜕𝑓⁄
( 𝜕𝑥)
− 𝜕𝑓 .
( ⁄𝜕𝑦)

(5)

Differentiation of composite functions:

Let z be function of x and y and that 𝑥 = 𝜑(𝑢, 𝑣) and 𝑦 = 𝜙(𝑢, 𝑣) are
functions of u and v then,

z f x f y &𝜕𝑧 = 𝜕𝑓 𝜕𝑥 + 𝜕𝑓 𝜕𝑦.
= + 𝜕𝑣 𝜕𝑥 𝜕𝑣 𝜕𝑦 𝜕𝑣
u x u y u
Similarly, if 𝑧 = 𝑓(𝑢, 𝑣) are functions of u and v and if𝑢 = 𝜑(𝑥, 𝑦)
and 𝑦 = 𝜙(𝑥, 𝑦) are functions of x and y then,
z f u f v & 𝜕𝑧 = 𝜕𝑓 𝜕𝑢 + 𝜕𝑓 𝜕𝑣 .
= + 𝜕𝑦 𝜕𝑢 𝜕𝑦 𝜕𝑣 𝜕𝑦
x u x v x
Note:
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
16
 Department of Mathematics

1) The above formulae can be extended to functions of three or more

variables and expressions are called Chain rule for partial
differentiation.

2) The second and higher order partial derivatives of 𝑧 = 𝑓(𝑥, 𝑦) can

be obtained by repeated applications of the above expressions.

Examples:

1. Find the total differential of

(i) 𝑒 𝑥 [𝑥 𝑠𝑖𝑛 𝑦 + 𝑦 𝑐𝑜𝑠 𝑦] (ii)𝑒 𝑥𝑦𝑧
Solution:
(i) Let 𝑧 = 𝑓(𝑥, 𝑦) = 𝑒 𝑥 [𝑥 𝑠𝑖𝑛 𝑦 + 𝑦 𝑐𝑜𝑠 𝑦] then
𝜕𝑧
= 𝑒 𝑥 [(1 + 𝑥) 𝑠𝑖𝑛 𝑦 + 𝑦 𝑐𝑜𝑠 𝑦]
𝜕𝑥
𝜕𝑧
and𝜕𝑦 = 𝑒 𝑥 [(1 + 𝑥) 𝑐𝑜𝑠 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦]
𝜕𝑧 𝜕𝑧
Hence, using formula (1), 𝑑𝑧 = 𝜕𝑥 𝑑𝑥 + 𝜕𝑦 𝑑𝑦 = 𝑒 𝑥 [(1 +
𝑥) 𝑠𝑖𝑛 𝑦 + 𝑦 𝑐𝑜𝑠 𝑦]𝑑𝑥 + 𝑒 𝑥 [(1 + 𝑥) 𝑐𝑜𝑠 𝑦 − 𝑦 𝑠𝑖𝑛 𝑦]𝑑𝑦

(ii) Let 𝑢 = 𝑓(𝑥, 𝑦, 𝑧) = 𝑒 𝑥𝑦𝑧 Then

𝜕𝑢 𝜕𝑢 𝜕𝑢
= (𝑦𝑧)𝑒 𝑥𝑦𝑧 ;𝜕𝑦 = (𝑥𝑧)𝑒 𝑥𝑦𝑧 ; 𝜕𝑧 = (𝑥𝑦)𝑒 𝑥𝑦𝑧
𝜕𝑥
∴Total differential of 𝑧 = 𝑓(𝑥, 𝑦, 𝑧) is (see formula (3) above)
𝜕𝑢 𝜕𝑢 𝜕𝑢
𝑑𝑢 = 𝑑𝑥 + 𝑑𝑦 + 𝑑𝑧 = 𝑒 𝑥𝑦𝑧 (𝑦𝑧𝑑𝑥 + 𝑧𝑥𝑑𝑦 + 𝑥𝑦𝑑𝑧)
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑑𝑧
2. Find 𝑑𝑡 if
(i) 𝑧 = 𝑥𝑦 2 + 𝑥 2 𝑦,where 𝑥 = 𝑎𝑡 2 ,𝑦 = 2𝑎𝑡
𝑦
(ii) 𝑧 = 𝑡𝑎𝑛−1( ⁄𝑥),where𝑥 = 𝑒 𝑡 − 𝑒 −𝑡 ,𝑦 = 𝑒 𝑡 + 𝑒 −𝑡
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
17
 Department of Mathematics

Solution:
(i) Consider 𝑧 = 𝑥𝑦 2 + 𝑥 2 𝑦
𝜕𝑧 𝜕𝑧
= 𝑦 2 + 2𝑥𝑦&𝜕𝑦 = 2𝑥𝑦 + 𝑥 2
𝜕𝑥
𝑑𝑥 𝑑𝑦
Since 𝑥 = 𝑎𝑡 2 &𝑦 = 2𝑎𝑡, = 2𝑎𝑡& 𝑑𝑡 = 2𝑎
𝑑𝑡
Hence, using formula (2),
𝑑𝑧 𝜕𝑧 𝑑𝑥 𝜕𝑧 𝑑𝑦
= + = (𝑦 2 + 2𝑥𝑦)(2𝑎𝑡) + (2𝑥𝑦 + 𝑥 2 )(2𝑎)
𝑑𝑡 𝜕𝑥 𝑑𝑡 𝜕𝑦 𝑑𝑡
𝑑𝑧
To get ( 𝑑𝑡 ) explicitly in terms of t, substitute
𝑑𝑧
𝑥 = 𝑎𝑡 2 &𝑦 = 2𝑎𝑡, to get (𝑑𝑡 ) = (4𝑎2 𝑡 2 + 4𝑎2 𝑡 3 )2𝑎𝑡 +
(2𝑎𝑡 2 × 2𝑎𝑡 + 𝑎2 𝑡 4 )2𝑎 = 2𝑎3 (8𝑡 3 + 5𝑡 4 )

𝑦
(ii) Consider𝑧 = 𝑡𝑎𝑛−1( ⁄𝑥)
𝜕𝑧 −𝑦 𝜕𝑧 𝑥
= 𝑥 2 +𝑦 2,𝜕𝑦 = 𝑥 2 +𝑦 2
𝜕𝑥
Since 𝑥 = 𝑒 𝑡 − 𝑒 −𝑡 &𝑦 = 𝑒 𝑡 + 𝑒 −𝑡 ,
dx 𝑑𝑦
= e t + e −t = y 𝑑𝑡 = 𝑒 𝑡 − 𝑒 −𝑡 = 𝑥
dt
𝑑𝑧 𝜕𝑧 𝑑𝑥 𝜕𝑧 𝑑𝑦
Hence 𝑑𝑡 = 𝜕𝑥 𝑑𝑡 + 𝜕𝑦 𝑑𝑡 (see equation (2))
−𝑦 𝑥 𝑥2 − 𝑦2
=( 2 ) 𝑦 + ( ) 𝑥 = ( )
𝑥 + 𝑦2 𝑥2 + 𝑦2 𝑥2 + 𝑦2
Substituting 𝑥 = 𝑒 𝑡 − 𝑒 −𝑡 &𝑦 = 𝑒 𝑡 + 𝑒 −𝑡 ,
𝑑𝑧 −2
= 2𝑡
𝑑𝑡 𝑒 + 𝑒 −2𝑡
𝒅𝒚
3. Find (𝒅𝒙) if
(i) 𝒙𝒚 + 𝒚𝒙 =c
(ii) 𝑥 + 𝑒 𝑦 = 2𝑥𝑦

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

18
 Department of Mathematics

Solution:
(i) Let 𝑧 = 𝑓(𝑥, 𝑦) = 𝑥 𝑦 + 𝑦 𝑥 = Constant.
using formula (5)
𝜕𝑓⁄
𝑑𝑦 𝜕𝑥
=−
𝑑𝑥 𝜕𝑓
⁄𝜕𝑦

𝜕𝑓 𝜕𝑓
But 𝜕𝑥 = 𝑦𝑥 𝑦−1 + 𝑦 𝑥 𝑙𝑜𝑔 𝑦 and 𝜕𝑦 = 𝑥 𝑦 𝑙𝑜𝑔 𝑥 + 𝑥𝑦 𝑥−1
putting these in (1),
𝑑𝑦 𝑦𝑥 𝑦−1 + 𝑦 𝑥 𝑙𝑜𝑔 𝑦
= −{ 𝑦 }
𝑑𝑥 𝑥 𝑙𝑜𝑔 𝑥 + 𝑥𝑦 𝑥−1

(ii) Let 𝑧 = 𝑓(𝑥, 𝑦) = 𝑒 𝑥 + 𝑒 𝑦 − 2𝑥𝑦=Constant

𝜕𝑓 𝜕𝑓
Now, 𝜕𝑥 = 𝑒 𝑥 − 2𝑦and𝜕𝑦 = 𝑒 𝑦 − 2𝑥 Using this in (1),
𝜕𝑓⁄
𝑑𝑦 𝜕𝑥 𝑒 𝑥 − 2𝑦
= −{ } = −{ 𝑦 }
𝑑𝑥 𝜕𝑓 𝑒 − 2𝑥
⁄𝜕𝑦

4. (i) If 𝑧 = 𝑓(𝑥, 𝑦),where 𝑥 = 𝑟 𝑐𝑜𝑠 𝜃,𝑦 = 𝑟 𝑠𝑖𝑛 𝜃 show that

𝜕𝑧 2 𝜕𝑧 2 𝜕𝑧 2 1 𝜕𝑧 2
( ) +( ) = ( ) + 2( )
𝜕𝑥 𝜕𝑦 𝜕𝑟 𝑟 𝜕𝜃
(ii) If 𝑧 = 𝑓(𝑥, 𝑦),where𝑥 = 𝑒 + 𝑒 &𝑦 = 𝑒 −𝑢 − 𝑒 𝑣 , show that
𝑢 −𝑣

𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
− =𝑥 −𝑦
𝜕𝑢 𝜕𝑣 𝜕𝑥 𝜕𝑦
Solution:
(i) As 𝑥 = 𝑟 𝑐𝑜𝑠 𝜃 and𝑦 = 𝑟 𝑠𝑖𝑛 𝜃,
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
= 𝑐𝑜𝑠 𝜃,𝜕𝜃 = −𝑟 𝑠𝑖𝑛 𝜃; 𝜕𝑟 = 𝑠𝑖𝑛 𝜃&𝜕𝜃 = 𝑟 𝑐𝑜𝑠 𝜃.
𝜕𝑟
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
19
 Department of Mathematics

Using chain rule,

𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑧
( )= + = (𝑐𝑜𝑠 𝜃) + (𝑠𝑖𝑛 𝜃)
𝜕𝑟 𝜕𝑥 𝜕𝑟 𝜕𝑦 𝜕𝑟 𝜕𝑥 𝜕𝑦

 z  z x z y z z
 = + = (− r sin  ) + (r cos )
   x  y  x y

Squaring on both sides and adding the above equations,

𝜕𝑧 2 𝜕𝑧 2 𝜕𝑧 2 𝜕𝑧 𝜕𝑧
( ) = ( ) 𝑐𝑜𝑠 2 𝜃 + ( ) 𝑠𝑖𝑛2 𝜃 + 2 ( ) ( ) 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜃
𝜕𝑟 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦
2 2
1 𝜕𝑧 𝜕𝑧 𝜕𝑧 2
( ) = ( ) 𝑐𝑜𝑠 𝜃 + ( ) 𝑠𝑖𝑛2 𝜃
2
𝑟 2 𝜕𝜃 𝜕𝑥 𝜕𝑦
𝜕𝑧 𝜕𝑧
− 2 ( ) ( ) 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜃
𝜕𝑥 𝜕𝑦
Adding the above equations,
𝜕𝑧 2 1 𝜕𝑧 2 𝜕𝑧 2 𝜕𝑧 2
( ) + 2 ( ) = {( ) + ( ) } (𝑐𝑜𝑠 2 𝜃 + 𝑠𝑖𝑛2 𝜃)
𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑥 𝜕𝑦
𝜕𝑧 2 𝜕𝑧 2
=(𝜕𝑥) + (𝜕𝑦) .
(ii) Given 𝑥 = 𝑒 𝑢 + 𝑒 −𝑣 &𝑦 = 𝑒 −𝑢 − 𝑒 𝑣 ,
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
= 𝑒 𝑢 ,𝜕𝑣 = −𝑒 −𝑣 ,𝜕𝑢 = −𝑒 −𝑢 &𝜕𝑣 = −𝑒 𝑣
𝜕𝑢
Using Chain rule (6)
𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝑢 𝜕𝑧
( )= + = (𝑒 ) + (−𝑒 −𝑢 )
𝜕𝑢 𝜕𝑥 𝜕𝑢 𝜕𝑦 𝜕𝑢 𝜕𝑥 𝜕𝑦
𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧 𝜕𝑧
( )= + = (−𝑒 −𝑣 ) + (−𝑒 𝑣 )
𝜕𝑣 𝜕𝑥 𝜕𝑣 𝜕𝑦 𝜕𝑣 𝜕𝑥 𝜕𝑦
𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
 ( ) − ( ) = (𝑒 𝑢 + 𝑒 −𝑣 ) − (𝑒 −𝑢 − 𝑒 𝑣 )
𝜕𝑢 𝜕𝑣 𝜕𝑥 𝜕𝑦

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

20
 Department of Mathematics

𝜕𝑧 𝜕𝑧
= 𝑥− 𝑦
𝜕𝑥 𝜕𝑦
𝝏𝒖 𝝏𝒖 𝝏𝒖
5. (i) If 𝒖 = 𝒇(𝒙𝒛, 𝒚/𝒛) then show that𝒙 𝝏𝒙 − 𝒚 𝝏𝒚 − 𝒛 𝝏𝒛 = 𝟎
𝜕𝐻 𝜕𝐻 𝜕𝐻
(ii) If 𝐻 = 𝑓(𝑥 − 𝑦, 𝑦 − 𝑧, 𝑧 − 𝑥), show that 𝜕𝑥 + 𝜕𝑦 + =
𝜕𝑧
0Solution:
𝑦
(i) Let 𝑢 = 𝑓(𝑣, 𝑤), where 𝑣 = 𝑥𝑧 and𝑤 = ⁄𝑧
𝜕𝑣
= 𝑧,
𝜕𝑣
= 0,
𝜕𝑣
= 𝑥 &
𝜕𝑤
= 0,
𝜕𝑤
= 1⁄ ,𝜕𝑤 = −𝑦⁄
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝑧 𝜕𝑧 𝑧2
Using chain rule,
𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑤 𝜕𝑢 𝜕𝑢 𝜕𝑢
= + = (𝑧) + (0) = 𝑧
𝜕𝑥 𝜕𝑣 𝜕𝑥 𝜕𝑤 𝜕𝑥 𝜕𝑣 𝜕𝑤 𝜕𝑣

𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑤 𝜕𝑢 𝜕𝑢 1 1 𝜕𝑢
= + = (0) + ( ⁄𝑧) =
𝜕𝑦 𝜕𝑣 𝜕𝑦 𝜕𝑤 𝜕𝑦 𝜕𝑣 𝜕𝑤 𝑧 𝜕𝑤
𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑤 𝜕𝑢 𝜕𝑢 −𝑦 𝜕𝑢 𝑦 𝜕𝑢
= + = (𝑥) + ( ⁄𝑧 2 ) = 𝑥 −
𝜕𝑧 𝜕𝑣 𝜕𝑧 𝜕𝑤 𝜕𝑧 𝜕𝑣 𝜕𝑤 𝜕𝑣 𝑧 2 𝜕𝑤
From these,
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑢 𝑦 𝜕𝑢 𝜕𝑢 𝑦 𝜕𝑢
𝑥 𝜕𝑥 − 𝑦 𝜕𝑦 − 𝑧 𝜕𝑧 = 𝑥𝑧 𝜕𝑣 − 𝑧 𝜕𝑤 − 𝑧 (𝑥 𝜕𝑣 − 𝑧 2 𝜕𝑤)= 0

(ii) Let 𝐻 = 𝑓(𝑢, 𝑣, 𝑤) Where 𝑢 = 𝑥 − 𝑦, 𝑣 = 𝑦 − 𝑧, 𝑤 = 𝑧 − 𝑥

𝜕𝑢 𝜕𝑢 𝜕𝑢
Now 𝜕𝑥 = 1, 𝜕𝑦 = −1, 𝜕𝑧 = 0
𝜕𝑣 𝜕𝑣 𝜕𝑣
= 0, = 1, = −1
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑤 𝜕𝑤 𝜕𝑢
= −1, = 0, =1
𝜕𝑥 𝜕𝑦 𝜕𝑧
Using Chain rule,
𝜕𝐻 𝜕𝐻 𝜕𝑢 𝜕𝐻 𝜕𝑣 𝜕𝐻 𝜕𝑤 𝜕𝐻 𝜕𝐻 𝜕𝐻
= + + = (1) + (0) + (−1)
𝜕𝑥 𝜕𝑢 𝜕𝑥 𝜕𝑣 𝜕𝑥 𝜕𝑤 𝜕𝑥 𝜕𝑢 𝜕𝑣 𝜕𝑤
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
21
 Department of Mathematics

𝜕𝐻 𝜕𝐻 𝜕𝑢 𝜕𝐻 𝜕𝑣 𝜕𝐻 𝜕𝑤 𝜕𝐻 𝜕𝐻 𝜕𝑣
= + + = (−1) + (1) + (0)
𝜕𝑦 𝜕𝑢 𝜕𝑦 𝜕𝑣 𝜕𝑦 𝜕𝑤 𝜕𝑦 𝜕𝑢 𝜕𝑣 𝜕𝑤
𝜕𝐻 𝜕𝐻 𝜕𝑢 𝜕𝐻 𝜕𝑣 𝜕𝐻 𝜕𝑤 𝜕𝐻 𝜕𝐻 𝜕𝐻
= + + = (0) + (−1) + (1)
𝜕𝑧 𝜕𝑢 𝜕𝑧 𝜕𝑣 𝜕𝑧 𝜕𝑤 𝜕𝑧 𝜕𝑢 𝜕𝑣 𝜕𝑤
Adding the above equations,
𝜕𝐻 𝜕𝐻 𝜕𝐻
+ 𝜕𝑦 + = 0.
𝜕𝑥 𝜕𝑧
6. The length and breadth of a rectangle are increasing at the rate of
1.5cm/sec and 0.5 cm/sec respectively. Find the rate at which the
area is increasing at the instant when length is 40 cm and breadth
is 30 cm respectively.
Solution: Let 𝑥 be the length, 𝑦 be the breadth, 𝑠 be the area at any
time 𝑡
So 𝑠 = 𝑥𝑦
𝑑𝑠
It is required to find 𝑑𝑡 when 𝑥 = 40𝑐𝑚and 𝑦 = 30𝑐𝑚, given that
dx 𝑑𝑦
= 1.5 cm / sec 𝑑𝑡 = 0.5𝑐𝑚/𝑠𝑒𝑐
dt
𝑑𝑠 𝜕𝑠 𝑑𝑥 𝜕𝑠 𝑑𝑦
= +
𝑑𝑡 𝜕𝑥 𝑑𝑡 𝜕𝑦 𝑑𝑡
𝑑𝑥 𝑑𝑦
=𝑦 +𝑥 = 30 × 1.5 + 40 × 0.5 = 65
𝑑𝑡 𝑑𝑡
Therefore increasing at the rate 65𝑐𝑚2 /𝑠𝑒𝑐.
7. If 𝑧 = 𝑓(𝑢, 𝑣) and 𝑢 = 𝑎𝑥 + 𝑏𝑦, 𝑣 = 𝑎𝑦 − 𝑏𝑥, show that

𝝏𝟐 𝒛 𝝏𝟐 𝒛 𝟐 𝟐
𝝏𝟐 𝒛 𝝏𝟐 𝒛
+ = (𝒂 + 𝒃 ) ( 𝟐 + 𝟐 )
𝝏𝒙𝟐 𝝏𝒚𝟐 𝝏𝒖 𝝏𝒗
𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣
Proof:𝜕𝑥 = 𝑎, 𝜕𝑦 = 𝑏, 𝜕𝑥 = −𝑏, 𝜕𝑦 = 𝑎
𝜕𝑧 𝜕𝑧 𝜕𝑢 𝜕𝑧 𝜕𝑣 𝜕𝑧 𝜕𝑧
= + = 𝑎− 𝑏
𝜕𝑥 𝜕𝑢 𝜕𝑥 𝜕𝑣 𝜕𝑥 𝜕𝑢 𝜕𝑣

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

22
 Department of Mathematics

Apply chain rule 𝑧 → 𝜕𝑧⁄𝜕𝑥

𝜕 𝜕𝑧 𝜕 𝜕𝑧 𝜕𝑢 𝜕 𝜕𝑧 𝜕𝑣
( )= ( )⋅ + ( )⋅
𝜕𝑥 𝜕𝑥 𝜕𝑢 𝜕𝑥 𝜕𝑥 𝜕𝑣 𝜕𝑥 𝜕𝑥
𝜕 𝜕𝑧 𝜕𝑧 𝜕 𝜕𝑧 𝜕𝑧
= (𝑎 −𝑏 )⋅𝑎+ (𝑎 − 𝑏 ) ⋅ (−𝑏)
𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣 𝜕𝑢 𝜕𝑣
2 2 2 2
𝜕 𝑧 2
𝜕 𝑧 𝜕 𝜕 2
𝜕 2𝑧
= 𝑎 − 𝑎𝑏 − 𝑎𝑏 + 𝑏
𝜕𝑥 2 𝜕𝑢2 𝜕𝑢𝜕𝑣 𝜕𝑣𝜕𝑢 𝜕𝑣 2
𝜕2𝑧 𝜕2 𝑧 𝜕2 𝜕2 𝜕2 𝑧
Similarly, 𝜕𝑦 2 = 𝑏 2 𝜕𝑢2 + 𝑎𝑏 𝜕𝑢𝜕𝑣 + 𝑎𝑏 𝜕𝑣𝜕𝑢 + 𝑎2 𝜕𝑣2
𝜕2𝑧 𝜕2 𝑧
𝐿𝐻𝑆 = (𝑎2 + 𝑏 2 ) ( + )
𝜕𝑢2 𝜕𝑣 2
Exercise:
1. Find the total differentials of
(i) 𝑥𝑦𝑧 + (𝑥𝑦𝑧)−1 (ii) 𝑥 2 𝑦 + 𝑦 2 𝑧 + 𝑧 2 𝑥
1 2
ANS: (i) [1 − (𝑥𝑦𝑧)2 ] (𝑦𝑧𝑑𝑥 + 𝑧𝑥𝑑𝑦 + 𝑥𝑦𝑑𝑧), (ii) (2𝑥𝑦 + 𝑧 )𝑑𝑥 +
(2𝑦𝑧 + 𝑥 2 )𝑑𝑦 + (𝑦 2 + 2𝑥𝑧)
𝑑𝑢
2. Find ( 𝑑𝑡 ) If
(i)𝑢 = 𝑥 2 − 𝑦 2, 𝑥 = 𝑒 𝑡 𝑐𝑜𝑠 𝑡 , 𝑦 = 𝑒 𝑡 𝑠𝑖𝑛 𝑡
(ii) 𝑢 = 𝑠𝑖𝑛 𝑥 𝑦 2 , 𝑥 = 𝑙𝑜𝑔 𝑡 , 𝑦 = 𝑒 𝑡
(iii) 𝑢 = 𝑥𝑦 + 𝑦𝑧 + 𝑧𝑥, 𝑥 = 1⁄𝑡 , 𝑦 = 𝑒 𝑡 , 𝑧 = 𝑒 −𝑡
(iv)𝑢 = 𝑙𝑜𝑔(𝑥 + 𝑦 + 𝑧) , 𝑥 = 𝑒 −𝑡 , 𝑦 = 𝑠𝑖𝑛 𝑡 , 𝑧 = 𝑐𝑜𝑠 𝑡
2𝑡 1 2𝑡 −(𝑒 𝑡 +𝑒 −𝑡 ) 2
ANS: (i) 2e , (ii) 𝑒 [ 𝑡 + 2 𝑙𝑜𝑔 𝑡] 𝑐𝑜𝑠( 𝑒 𝑙𝑜𝑔 𝑡),(iii)
2t
𝑡2
+ ,
𝑡
(𝑐𝑜𝑠 𝑡−𝑠𝑖𝑛 𝑡−𝑒 −𝑡 )
(iv) (𝑐𝑜𝑠 𝑡+𝑠𝑖𝑛 𝑡+𝑒 −𝑡 )
𝑑𝑦
3. Find (𝑑𝑥 ) each of the following cases:-
(i) 𝑥 𝑠𝑖𝑛(𝑥 − 𝑦) = (𝑥 + 𝑦) (ii) (𝑐𝑜𝑠 𝑥)𝑦 = (𝑠𝑖𝑛 𝑦)𝑥
(𝑥 𝑐𝑜𝑠(𝑥−𝑦)+𝑠𝑖𝑛(𝑥−𝑦)−1) 𝑦 𝑡𝑎𝑛 𝑥+𝑙𝑜𝑔𝑠𝑖𝑛 𝑦
ANS: (i) , (ii) 𝑙𝑜𝑔𝑐𝑜𝑠 𝑥−𝑥 𝑐𝑜𝑡 𝑦
𝑥 𝑐𝑜𝑠(𝑥−𝑦)+1

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

23
 Department of Mathematics

4. If 𝑧 = 𝑓(𝑥, 𝑦) and 𝑥 = 𝑢 − 𝑣, 𝑦 = 𝑢𝑣 Prove that

𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧 𝜕𝑧
(i) (𝑢 + 𝑣) 𝜕𝑥 = 𝑢 𝜕𝑥 − 𝑣 𝜕𝑣 (ii) (𝑢 + 𝑣) 𝜕𝑦 = 𝜕𝑢 + 𝜕𝑣
5. If 𝑢 = 𝑥 2 − 𝑦 2 , 𝑧 = 2𝑟 − 3𝑠 + 4, 𝑦 = −𝑟 + 8𝑠 − 5, Prove that
𝜕𝑢
= 4𝑥 + 2𝑦
𝜕𝑟
𝜕𝑢
6. If 𝑧 = 𝑓(𝑟, 𝑠), where 𝑟 = 𝑥 + 𝑎𝑡, 𝑦 = 𝑦 + 𝑏𝑡 Prove that =
𝜕𝑡
𝜕𝑢 𝜕𝑢
𝑎 𝜕𝑥 + 𝑏 𝜕𝑦.
𝜕2𝑧 𝜕2𝑧
7. If 𝑧 = 𝑓(𝑢, 𝑣), 𝑢 = 𝑥 2 − 𝑦 2 , 𝑣 = 2𝑥𝑦 Show that + 𝜕𝑦 2 =
𝜕𝑥 2
𝜕2𝑧 𝜕2𝑧
4(𝑥 2 + 𝑦 2 ) (𝜕𝑢2 + 𝜕𝑣 2 )
If 𝑢 = 𝑓(𝑥, 𝑦), 𝑥 + 𝑦 = 2𝑒 𝜃 𝑐𝑜𝑠 𝜑 , 𝑥 − 𝑦 = 2𝑖𝑒 𝜃 𝑠𝑖𝑛 𝜑 prove that
𝜕2𝑢 𝜕2𝑢 𝜕2𝑧
+ 𝜕𝜑2 = 4𝑥𝑦 𝜕𝑥𝜕𝑦.
𝜕𝜃2

Jacobians:

Carl Gustav Jacob Jacobi; 10

December 1804 – 18 February
1851) was a German
mathematician, who made
fundamental contributions to
elliptic functions, dynamics,
differential equations, and
number theory.

If ‘u’ and ‘v’ are functions of variables ‘x’ and ‘y’, then the
determinant

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

24
 Department of Mathematics

𝑢𝑥 𝑢𝑦
|𝑣 𝑣𝑦 | is called the Jacobian of u, v with respect to x, y and denoted
𝑥
𝜕(𝑢,𝑣)
by 𝜕(𝑥,𝑦). Likewise, it is possible to define Jacobian of more variables
in similar manner. These have an important application in coordinate
transformation from one system to another.

Note:
𝜕(𝑢,𝑣) 𝜕(𝑥,𝑦)
1. If J =𝜕(𝑥,𝑦) ≠ 0, then exists and is known as inverse of ‘J’
𝜕(𝑢,𝑣)
with JJ '=1.
2. 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)
𝜕(𝑢,𝑣)
3. If J =𝜕(𝑥,𝑦)= 0, then u,v are said to be functionally dependent. In
this case u can be expressed as a function of v and vice versa or
functional dependence can be expressed as f(u,v) = 0.

Examples:
𝜕(𝑥,𝑦)
1. If x = r cosθ, y = r sinθ find𝜕(𝑟,𝜃).
Solution: By definition we have
xr x cos − r sin 
𝜕(𝑥,𝑦)
J =𝜕(𝑟,𝜃)= = =r
yr y sin  r cos
The above are the transformations between Cartesian and polar
𝑠𝑐𝑑𝑠𝑑𝑎𝑎𝑎𝑎𝑣𝑣𝑣
coordinates. The inverse
2 2 −1 𝑦
transforms are 𝑟 = √𝑥 + 𝑦 and 𝜃 = tan (𝑥 ).

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

25
 Department of Mathematics

𝜕(r,θ)
It can be verified that the Jacobian J1 = of these functions is 1/r and
𝜕(x,y)
therefore JJ1 =1.
2. If 𝑥 = 𝑟 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜙 𝑦 = 𝑟 𝑠𝑖𝑛 𝜃 𝑠𝑖𝑛 𝜙 𝑧 = 𝑟 𝑐𝑜𝑠 𝜃 find
𝜕(𝑥,𝑦,𝑧)
𝜕(𝑟,𝜃,𝜙)

𝑥𝑟 𝑥𝜃 𝑥𝜙
𝜕(𝑥,𝑦,𝑧)
Solution:J=𝜕(𝑟,𝜃,𝜙)=| 𝑦𝑟 𝑦𝜃 𝑦𝜙 |=
𝑧𝑟 𝑧𝜃 𝑧𝜙

𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜙 𝑟 𝑐𝑜𝑠 𝜃 𝑐𝑜𝑠 𝜙 −𝑟 𝑠𝑖𝑛 𝜃 𝑠𝑖𝑛 𝜙

| 𝑠𝑖𝑛 𝜃 𝑠𝑖𝑛 𝜙 𝑟 𝑐𝑜𝑠 𝜃 𝑠𝑖𝑛 𝜙 𝑟 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜙 |
𝑐𝑜𝑠 𝜃 −𝑟 𝑠𝑖𝑛 𝜃 0

On expanding by the last row we get, J = 𝑟 2 𝑠𝑖𝑛 𝜃 (𝑐𝑜𝑠 2 𝜃 + 𝑠𝑖𝑛2 𝜃)

= 𝑟 2 𝑠𝑖𝑛 𝜃

The above are the transformations between Cartesian and spherical

2 2 2
coordinates. The inverse transforms are 𝑟 = √𝑥 + 𝑦 + 𝑧 , 𝜃 =
𝑥 2 +𝑦 2 𝑦
tan−1 (√ 𝑧2
)and 𝜑 = tan−1 (𝑥 ).
𝜕(r,θ, φ)
It can be easily verified that the Jacobian J1= of these functions is
𝜕(x,y,z)
1/r2sinθ and therefore JJ1 =1.

𝜕(𝑢,𝑣,𝑤)
3.If𝑢 = 𝑥 + 3𝑦 2 − 𝑧 3 , 𝑣 = 4𝑥 2 𝑦𝑧, 𝑤 = 2𝑧 2 − 𝑥𝑦 find at
𝜕(𝑥,𝑦,𝑧)
(1, -1, 0).

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

26
 Department of Mathematics

1 6𝑦 −3𝑧 2
𝜕(𝑢,𝑣,𝑤)
Solution: J= 𝜕(𝑥,𝑦,𝑧) = |8𝑥𝑦𝑧 4𝑥 2𝑧 4𝑥 2 𝑦 |at (1,-1, 0)
−𝑦 −𝑥 4𝑧
1 −6 0
J=|0 0 −4|=20
1 −1 0

4. If x = u(1-v), y = uv then compute J and J1 and verify that J J1=1.

𝜕(x, y) 𝑥𝑢 𝑥𝑣 1 − 𝑣 −𝑢
Solution: J = 𝜕(u,v) = |𝑦 𝑦𝑣 |=| 𝑣 | = u(1-v) + uv = u.
𝑢 𝑢

To find J1, express u and v in terms of x and y.

𝑦
Given x = u-uv → x= u-y → u = x + y, 𝑣 = x+y
𝜕(u, v) 𝑢𝑥 𝑢𝑦 1 1
J1 = 𝜕(x, y) = | 𝑣 |=| −𝑦 𝑥 | = 1 = 1.
𝑥 𝑣𝑦 (𝑥+𝑦) 2 (𝑥+𝑦)2 𝑥+𝑦 𝑢
1
J J =1.

𝑥+𝑦 𝜕(u, v)
5. If 𝑢 = 1−𝑥𝑦 and 𝑣 = 𝑡𝑎𝑛−1 𝑥 + 𝑡𝑎𝑛−1 𝑦 find 𝜕(x, y).

𝜕(u, v) 𝑢𝑥 𝑢𝑦
Solution: 𝜕(x, y) = | 𝑣 𝑣𝑦 |
𝑥

1+𝑦2 1+𝑥 2
(1−𝑥𝑦)2 (1−𝑥𝑦)2
=| 1 1
| =0
1+𝑥 2 1+𝑦 2

Since Jacobian is zero, u and v are functionally dependent.

Relation between u and v is tan 𝑣 = 𝑢.

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

27
 Department of Mathematics

Exercise:
𝑢,𝑣
1. .If 𝑢 + 𝑣 = 𝑒 𝑥 𝑐𝑜𝑠 𝑦 , 𝑢 − 𝑣 = 𝑒 𝑥 𝑠𝑖𝑛 𝑦, find 𝐽 (𝑥,𝑦).
𝑒𝑥 𝑒𝑥
[Hint: Start from 𝑢 = 2 (𝑐𝑜𝑠 𝑦 + 𝑠𝑖𝑛 𝑦) and𝑣 = 2 (𝑐𝑜𝑠 𝑦 − 𝑠𝑖𝑛 𝑦)
2. If 𝑢 = 𝑥𝑦𝑧, 𝑣 = 𝑥𝑦 + 𝑦𝑧 + 𝑧𝑥, 𝑤 = 𝑥 + 𝑦 + 𝑧, show that
𝜕(𝑢,𝑣,𝑤)
= (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥).
𝜕(𝑥,𝑦,𝑧)
3. If 𝑥 = 𝑒 𝑢 𝑠𝑒𝑐 𝑣 , 𝑦 = 𝑒 𝑢 𝑡𝑎𝑛 𝑣, show that
𝜕(𝑥,𝑦) 𝜕(𝑢,𝑣)
× 𝜕(𝑥,𝑦) = 1.
𝜕(𝑢,𝑣)
𝜕(𝑥,𝑦,𝑧)
4 .If 𝑥 + 𝑦 + 𝑧 = 𝑢, 𝑦 + 𝑧 = 𝑣, 𝑧 = 𝑢𝑣𝑤, find the value of .
𝜕(𝑢,𝑣,𝑤)

𝜕(𝑢,𝑣)
5. If 𝑢 = 2𝑥𝑦, 𝑣 = 𝑥 2 − 𝑦 2 , 𝑥 = 𝑟 𝑐𝑜𝑠 𝜃 , 𝑦 = 𝑟 𝑠𝑖𝑛 𝜃 , 𝑓𝑖𝑛𝑑 𝜕(𝑟,𝜃).
𝑒 2𝑥
Answers: 1. − , 4. u2v, 5. - 4r3.
2

Maxima and Minima of f(x, y):

In mathematics, the maximum and minimum (plural: maxima and

minima) of a function, known collectively as extrema (singular:
extremum), are the largest and smallest value that the function takes
at a point within a given neighborhood.

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

28
 Department of Mathematics

A function f (x, y) is said to have a Maximum value at (a,b) if there

exists a neighborhood point of (a,b) (say (a+h, b+k)) such that f (a,
b) >f (a+h, b+k).
Similarly, a function f (x, y) is said to have a Minimum value at (a,b)
if there exists a neighborhood point of (a,b) (say (a+h,b+k)) such that
f (a, b) <f (a+h, b+k).

Necessary condition:
Consider a function f = f(x, y) where x and y are independent real
variables. Suppose this function is continuous and has continuous
partial derivatives of second order in a region R in the x y – plane. Let
(a, b) be a point in the region. Then (a, b) is called a extremum point
of f(x, y), if f(x, b) is extremum at x=a and f(a, y) is extremum at y=b.
Since f(x, b) is a function of a single variable x, a necessary condition
𝜕𝑓
for the zero at x = a that is (𝑎, 𝑏) = 0Similarly a necessary
𝜕𝑥
𝜕𝑓
condition for f(a, y) to be extremum at y = b is that (𝑎, 𝑏) = 0.
𝜕𝑦
Thus, we have the following result.

A necessary condition for 𝑓(𝑥, 𝑦) to have an extremum at a point (a,

𝜕𝑓 𝜕𝑓
b) is that 𝜕𝑥 (𝑎, 𝑏) = 0and (𝑎, 𝑏) = 0.
𝜕𝑦
An extremum point of 𝑓(𝑥, 𝑦) is also called a stationary point or a
critical point.

Working procedure for finding Maxima and Minima of f(x,y):

Given f (x, y)

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

29
 Department of Mathematics

1. Find p, q, r, s, t,
2. Solve p = 0, q = 0 simultaneously
3. Let the points be (x1, y1), (x2, y2)….. .At each such point find r t
- s2.
4. At (x1, y1)
r t - s2 > 0, r > 0 → point of minima and f(x1, y1) is
minimum value of the function
r t - s2 > 0, r < 0 → point of maxima and f(x1, y1) is
maximum value of the function
r t - s2 < 0 → f(x1, y1) is neither maximum nor minimum and
(x1, y1) is called saddle point
r t - s2 = 0 → further investigation needed

Example
1. Show that z(x, y) = xy(a-x-y), a> 0, is maximum at the point
(a/3,a/3).
Solution: For the given function,
𝜕𝑓 𝜕𝑓
we have 𝑝 = 𝜕𝑥 = 𝑦(𝑎 − 2𝑥 − 𝑦)𝑎𝑛𝑑 𝑞 = 𝜕𝑦 = 𝑥(𝑎 − 𝑥 − 2𝑦)
Solving p = 0, q = 0, we get (a/3,a/3). Hence (a/3,a/3) is a critical
point.
𝜕 2𝑓 𝜕 2𝑓 𝜕 2𝑓
𝑟 = 2 =-2y, 𝑠 = =a-2x-2y, t= 2 = −2𝑥 𝑎𝑛𝑑 𝑟𝑡 − 𝑠 2
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦
2
= 4𝑥𝑦 − (𝑎 − 2𝑥 − 2𝑦)
At the point (a/3,a/3), r < 0, and rt - s2> 0.
It follows that f(x, y) is maximum at the point (a/3,a/3). The
maximum value is (1/27) a3.

2. Examine the following functions for extreme values f = x4 + y4 –

2x2 + 4xy – 2y2.
Solution: Given f = x4 + y4 – 2x2 + 4xy – 2y2
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
30
 Department of Mathematics

fx= 4x3 – 4x + 4y, fy = 4y3 – 4x – 4y

fxy = 4, fxx = 12 x2 – 4, fyy = 12y2 – 4
Solving fx= 0, fy = 0,
4x3 – 4x + 4y = 0 → (1) and 4y3 + 4x – 4y = 0 → (2)
Adding (1) & (2), we get 4 (x3 + y3) = 0 ⇒x3 + y3 = 0 so, y = - x
Substitute y = -x in (1), we get
4x3 – 4x – 4x = 0 ⇒ 4x3 – 8x = 0 ⇒ x3 – 2x = 0 ⇒ x (x2 – 2) = 0
⇒ x = 0 &x2 – 2 =0
⇒ x = 0 &x = ±√2 and y = 0 & y = ∓√2
∴ The critical points are (0,0), (+√2, -√2) and ( -√2 , +√2)

At (0,0):
r = fxx (0,0) = - 4, s = fxy (0,0) = 4, t = fyy(0,0) = -4 and rt - s2 = (-4)
(-4) - 16 = 0
⇒ further investigation needed.
At (±√𝟐,∓ √𝟐 ):
r = fxx(±√2,∓ √2 )=24– 4 = 20 , s = fxy(±√2,∓ √2 ) = 4>0 and t =
fyy(±√2,∓ √2 ) = 20
∴ s2 – rt = 16 – (20) (20) = 16 – 400 = -384 < 0
⇒The function is minimum at (±√2,∓ √2 ) and fmin = f(±√2,∓ √2 ) =
-8
3. Find Maximum and Minimum of 2(𝑥 2 − 𝑦 2 ) − 𝑥 4 + 𝑦 4 .

Solution: Let 𝑢 = 2(𝑥 2 − 𝑦 2 ) − 𝑥 4 + 𝑦 4

𝜕𝑢 𝜕𝑢
For the stationary points 𝜕𝑥 = 0 and = 0.
𝜕𝑦
𝑢𝑥 = 4𝑥 − 4𝑥 3 = 0 ⇒ 4𝑥(1 − 𝑥 2 ) = 0 ⇒ 𝑥 = 0, 1, −1
𝑢𝑦 − 4𝑦 + 4𝑦 3 = 0 ⇒ −4𝑦(1 − 𝑦 2 ) = 0 ⇒ 𝑦 = 0, 1, −1

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

31
 Department of Mathematics

Stationary points are

(0,0), (0,1), (0, −1), (1,0), (1,1), (1, −1), (−1,0), (−1,1) and
(−1, −1)
𝑟 = 𝑢𝑥𝑥 = 4 − 12𝑥 2 , 𝑠 = 𝑢𝑥𝑦 = 0, 𝑡 = 𝑢𝑦𝑦 = −4 + 12𝑦 2

Point 𝑟 𝑠 𝑡 𝑟𝑡 𝑟 Nature of
− 𝑠2 the point
(0,0) 4 0 −4 < 0 >0 𝑆𝑎𝑑𝑑𝑙𝑒 𝑝𝑡
(0,1) 4 0 8 >0 >0 𝑀𝑖𝑛 𝑝𝑡
(0, −1) 4 0 8 >0 >0 𝑀𝑖𝑛 𝑝𝑡
(1,0) −8 0 −4 >0 <0 𝑀𝑎𝑥 𝑝𝑡
(1,1) −8 0 8 <0 <0 𝑆𝑎𝑑𝑑𝑙𝑒 𝑝𝑡
(1, −1) −8 0 8 <0 <0 𝑆𝑎𝑑𝑑𝑙𝑒 𝑝𝑡
(−1,0) −8 0 −4 >0 <0 𝑀𝑎𝑥 𝑝𝑡
(−1, 1) −8 0 8 <0 <0 𝑆𝑎𝑑𝑑𝑙𝑒 𝑝𝑡
(−1, −1) −8 0 8 <0 <0 𝑆𝑎𝑑𝑑𝑙𝑒 𝑝𝑡

u is maximum at (1,0),(-1,0), Maximum value is 1. U is minimum at

(0,1),(0,-1), minimum value is -1
4. A rectangle box, open at the top to have a volume of 32 cubic units,
what must be the dimension so that total surface area of the box is a
minimum.
Solution: Let 𝑥, 𝑦 and 𝑧 be the dimension of the box. (Length, breadth
and height of the box) Then Volume 𝑉 = 𝑥𝑦𝑧
Surface area is 𝑠 = 𝑥𝑦 + 2𝑦𝑧 + 2𝑧𝑥 = 192
32
Given 𝑉 = 32 so that 𝑥𝑦𝑧 = 32 ⇒ 𝑧 = 𝑥𝑦
32 32 64 64
𝑠 = 𝑥𝑦 + 2𝑦 + 2𝑥 = 𝑥𝑦 + + = 𝑓(𝑥, 𝑦)
𝑥𝑦 𝑥𝑦 𝑥 𝑦

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

32
 Department of Mathematics

64 64 64
𝑓𝑥 = 𝑦 − 𝑥 2 , 𝑓𝑦 = 𝑥 − 𝑦 2 , 𝑓𝑥 = 0, 𝑓𝑦 = 0 ⇒ 𝑦 − 𝑥 2 = 0, 𝑥−
64
=0
𝑦2
𝑥 2 𝑦 = 64, 𝑦 2 𝑥 = 64
⇒ 𝑥 2 𝑦 = 𝑦 2 𝑥 ⇒ 𝑥 = 𝑦 ⇒ 𝑥 2 𝑦 = 𝑥 3 = 64 ⇒ 𝑥 = 𝑦 = 4
Therefore (4, 4) is critical point.
128 128
𝑟 = 𝑓𝑥𝑥 = 3 𝑠 = 𝑓𝑥𝑦 = 1 𝑡 = 𝑓𝑦𝑦 = 3
𝑥 𝑦
At (4, 4)𝑟 = 2, 𝑠 = 1, 𝑡 = 2
𝑟𝑡 − 𝑠 2 = 4 − 1 = 3 > 0, r > 0.Therefore (4, 4) is minimum point.
Minimum volume 𝑓(4, 4) = 48.
Then dimensions of box which make surface area minimum are 𝑥 =
4, 𝑦 = 4 and 𝑧 = 2
Exercise:

1) Find the extreme values of f = x3y2 (1 – x – y)

2) Examine the function f(x,y) = 1+ sin( x2 + y2) for extremum.
3) Find the maxima and minima of the functions f (x,y) = x3 + y3 –
3axy, a> 0 is a constant.
4) The temp T at any point (𝑥, 𝑦, 𝑧) in space is 𝑇 = 400𝑥𝑦𝑧 2 . Find
the highest temp on the surface of the unit sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 1.
ANS: (1) Maximum at (1/2,1/3), maximum value =1/432.
(2) Minimum at (0,0), minimum value =1
(3) Minimum at (a,a), minimum value = - a3
(4) 50

Lagrange Multiplier method or Constrained Extrema

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
33
 Department of Mathematics

In practical situations one often comes across problems of

determining extrema of functions of many variables with the variables
constrained by certain conditions. In other words one requires
maxima or minima of functions where the variables satisfy certain
equality or inequality conditions, called constraints.

This is a special method of obtaining stationary values of a function

of three independent variables when the variables are subject to some
conditions. Suppose it is required to find extrema of 𝑓(𝑥, 𝑦, 𝑧) subject
to the condition that 𝜑(𝑥, 𝑦, 𝑧) = 0

To find the values of x, y, z for which f(x, y, z) is stationary

(maximum or minimum):

1. Find an auxiliary function F(x, y, z) defined by 𝐹(𝑥, 𝑦, 𝑧) =

𝑓(𝑥, 𝑦, 𝑧) + 𝜆𝜑(𝑥, 𝑦, 𝑧), where 𝜆 is a constant.
𝜕𝐹 𝜕𝐹 𝜕𝐹
2. Determine the partial derivatives 𝜕𝑥 , 𝜕𝑦 , 𝜕𝑧 and equate each of
these to zero. Thus we obtain three equations of the form
𝐹1 (𝑥, 𝑦, 𝑧, 𝜆) = 0, 𝐹2 (𝑥, 𝑦, 𝑧, 𝜆) = 0, 𝐹3 (𝑥, 𝑦, 𝑧, 𝜆) = 0.
These equations together with the constrain gives a value for 𝜆. For
this value of 𝜆, find x, y, z. Then (x, y, z) is a stationary point and the
corresponding value is the stationary value. This method of obtaining
stationary values of f(x, y, z) is called the Lagrange’s method of
multipliers.

Examples:

1. The temperature at any point (𝑥, 𝑦, 𝑧) in space = 400𝑥𝑦 2 𝑧 . Find

the highest temperature at the surface of unit sphere 𝑥 2 + 𝑦 2 + 𝑧 2 =
1 using Lagrange’s method of undetermined multipliers.
Solution: Let 𝑓(𝑥, 𝑦, 𝑧) = 400𝑥𝑦 2 𝑧

∅(𝑥, 𝑦, 𝑧) = 𝑥 2 + 𝑦 2 + 𝑧 2 − 1
Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)
34
 Department of Mathematics

𝐹(𝑥, 𝑦, 𝑧) = 𝑓(𝑥, 𝑦, 𝑧) + 𝜆𝜑(𝑥, 𝑦, 𝑧)

𝐹(𝑥, 𝑦, 𝑧) = 400𝑥𝑦 2 𝑧 + 𝜆(𝑥 2 + 𝑦 2 + 𝑧 2 − 1) (1)
200𝑦 2 𝑧
𝐹𝑥 = 400𝑦 2 𝑧 + 2𝜆𝑥 = 0 ⟹ = −𝜆 (2)
𝑥
400𝑥𝑦𝑧
𝐹𝑦 = 800𝑥𝑦𝑧 + 2𝜆𝑦 = 0 ⟹ = −𝜆 (3)
𝑦
200𝑥𝑦 2
𝐹𝑧 = 400𝑥𝑦 2 + 2𝜆𝑧 = 0 ⟹ = −𝜆 (4)
𝑧

1 1 1
Solving (2),(3),(4), we get 𝑥 = ± , 𝑦 = ± ,𝑧 = ±
2 √2 2

1 1 1
The highest temperature is 𝑇 = 400𝑥𝑦 2 𝑧 = 400 × 2 × 2 × 2 = 50.

2. A tank is in the form of a rectangular box open at the top. Find its
dimensions if its inner surface area is 192 m2 and maximum capacity.

Solution: Let the length, breadth and height of the box be 𝑥, 𝑦, 𝑧

meters respectively. Then volume 𝑉 = 𝑥𝑦𝑧 . It is given that 𝑥𝑦 +
2𝑦𝑧 + 2𝑧𝑥 = 192.
To determine 𝑥, 𝑦, 𝑧 so that V, subject to the condition, is maximum.

𝑉 = 𝑥𝑦𝑧 + 𝜆(𝑥𝑦 + 2𝑦𝑧 + 2𝑧𝑥 − 192)

𝜕𝑉 𝜕𝑉 𝜕𝑉
At the stationary points, 𝜕𝑥 = 0, 𝜕𝑦 = 0, 𝜕𝑧 = 0.
𝑦𝑧 + 𝜆(𝑦 + 𝑧) = 0 (1)
𝑧𝑥 + 𝜆(𝑦 + 2𝑧) = 0 (2)

𝑥𝑦 + 𝜆(2𝑦 + 2𝑥) (3)

𝑥𝑦 + 2𝑦𝑧 + 2𝑧𝑥 = 192) (4)

Solving (1), (2), (3) we get 𝑥 = −4𝜆, 𝑦 = −4𝜆, 𝑧 = −2𝜆,
from (4), 𝜆 = −2, 𝑥 = 8, 𝑦 = 8, 𝑧 = 4; (𝜆 =
2 is not admissible because this give 𝑥, 𝑦, 𝑧 < 0. Thus, for 𝑉 to be

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

35
 Department of Mathematics

maximum, length = 8m, breadth = 8, and height = 4m with volume =

256m3.

Exercise:
1. A tent on a square base of side x has its four sides vertical, of height
y and top in the form of a regular pyramid of height h. If the capacity
of the tent is given, find x and y in terms of h in order that the canvas
required for its construction is to be a minimum using Lagrange’s
multiplier method.
2. Find the dimensions of the largest rectangular parallelepiped that
𝑥2 𝑦2 𝑧2
can be enclosed in the ellipsoid 𝑎2 + 𝑏2 + 𝑐 2 = 1 using
Lagrange’s multiplier method.
ℎ 𝑎 𝑏 𝑐
ANS: 1. 𝑥 = √5ℎ, 𝑦 = 2 2. 𝑥 = ,𝑦 = ,𝑧 = .
√3 √3 √3

Video Links:
1. Partial derivatives
https://www.youtube.com/watch?v=P43qNhCj1Gg
https://www.youtube.com/watch?v=Q8mbXy0oJj8

2. Jacobian
https://www.khanacademy.org/math/multivariable-
calculus/multivariable-derivatives/jacobian/v/the-jacobian-matrix

3. Multivariable Maxima and Minima

https://www.khanacademy.org/math/multivariable-
calculus/applications-of-multivariable-derivatives/optimizing-
multivariable-functions-videos/v/multivariable-maxima-and-minima

Disclaimer: The content provided is prepared by department of

Mathematics for the specified syllabus by using reference books
mentioned in the syllabus. This material is specifically for the use of
RVCE students and for education purpose only.

Fundamentals of Linear Algebra, Calculus and Numerical Methods (MAT211AT)

36