4-Partial Derivatives and Their Applications
Uploaded by
adam4-Partial Derivatives and Their Applications
Uploaded by
adam4 PARTIAL DERIVATIVES AND THEIR
A PPLICATIONS
aaaaa
4.1 INTRODUCTON: FUNCTIONS OF SEVERAL VARIABLES
So far, we had discussed functions of a single real variable defined by y = f (x). Here in this
chapter, we extend the concept of functions of two or more variables.
If to each pair (x, y) of values of certain parts of the xy-plane, x ∈ IR (: R : the set of real
numbers), y ∈ IR or (x, y) ∈ IR × IR, there corresponds real valued z in such a way that
z = f (x, y), x ∈ IR, y ∈ IR or (x, y) ∈ IR2. …(1)
wherein x, y are independent variables and z is the dependent variable.
We may extend this notation to a function of n independent variables as
z = f(x1, x2, … , xn), (x1, x2, …, xn) ∈ IRn …(2)
wherein (x1, x2, …, xn) are n independent variables and z is the dependent variable.
In actual life we frequently come across such functions for instance the area A of a rectangle
of length x(as one independent variable) and breadth y (as the 2nd independent variable)
given by A = xy. We can extend it to a function of three variables by taking 3rd independent
variable z, the height giving rise to volume of the parallelopiped as V = xyz and so on a
function of n dimension may be defined.
The function z = f (x, y) of two independent variables represents a surface in three
dimensional space referred to a set of co-ordinate axes x, y, z is completely analogous to a
real function of a single independent variable x, say, y = f (x) represented geometrically by a
curve in the xy-plane.
4.2 PARTIAL DERIVATIVES
Consider the function z = f(x, y) of two independent variables x and y and extend the concept
of ordinary derivative of the function of one variable to the function z = f(x, y) by keeping y
constant while taking derivative with respect to x and keeping x constant while taking
derivative with respect to y. This way we had two different derivatives and to distinguish
them one is called partial derivative of z(x, y) with respect to x denoted by one of the
∂z ∂f
symbols , , f (x, y), Dx f given by
∂x ∂x x
∂z f (x + δx, y) − f (x, y)
= Lt
∂x δx →0 δx
263
264 Engineering Mathematics through Applications
Likewise, 2nd is the partial derivative of z (x, y) with respect to y denoted by one of the
∂z ∂f
symbols ∂y , ∂y , fy(x, y), Dy f given by
∂z f (x, y + δy) − f (x, y)
= Lt
∂y δy →0 δy
∂f
Geometrically, is the slop of the tangent drawn to the curve of intersection of the
∂x
surface z = f(x, y) with a plane parallel to the plane y = 0.
∂f ∂f
Further, in general and ∂y are functions of both x and y and, therefore can be
∂x
differentiated for higher order derivatives with respect to x or y.
Thus
∂ ∂f ∂2 f ∂ ∂f ∂2 f
= 2 = fxx , = = fyx ,
∂x ∂x ∂x ∂x ∂y ∂x∂y
∂ ∂f ∂2 f ∂ ∂f ∂2 f
= = fyy , = = fxy ,
∂y ∂y ∂y2 ∂y ∂x ∂y∂x
Also, sometime we denote these derivatives as
∂z ∂z ∂2z ∂ 2z ∂2 z
= p, = q, = r, = s, =t
∂x ∂y ∂x2 ∂x∂y ∂y2
∂2 f ∂2 f
In all ordinary cases, it has been observed that ∂x∂y = ∂y∂x , meaning thereby that the
order of partial differentiation is immaterial.
∂z ∂z
Example 1: If z = eax + by f(ax – by), prove that b ∂x + a ∂y = 2 abz [VTU, 2004]
Solution:
∂ z ( ax + by)
Here =e · a f (ax – by)+ e( ax + by) · f'(ax − by)· a
∂x
∂z
Similarly, = e(ax + by)· b · f (ax – by)+ e(ax +by) f'(ax − by)(−b)
∂y
∴
∂z ∂z
{
b ∂x + a ∂y = b e
(ax + by)
}
· a f (ax − by) + e(ax +by) f'(ax − by) a
+ {a[e(ax + by) bf(ax – by) + e(ax+ by) f'(ax – by)(– b)]}
= ab e(ax + by) f(ax – by) + ab e(ax + by) · f(ax – by)
= 2abz–
Example 2: Let r2 = x2 + y2 + z2 and V = r m, prove that Vxx + Vyy + Vzz = m(m + 1)r m – 2
[Raipur, 2005; PTU, 2006]
Partial Derivatives and their Applications 265
m
Solution: Given V = rm = (r2 )m/2 = (x2 + y2 + z2 ) 2 …(1)
Here Vxx denotes 2nd order partial derivative of V(x, y, z) with respect to x keeping y and z
constant.
∂ ∂ m
m m
−1
Thus Vx = V(x, y, z)= (x2 + y2 + z2 ) 2 = (x2 + y2 + z2 ) 2 2x
∂x ∂x 2
m− 2
= m x (x2 + y2 + z2 ) 2 …(2)
∂ m− 2
∂ m− 2 m− 2
and Vxx = mx(x2 + y2 + z2 ) 2 = m x ⋅ (x2 + y2 + z2 ) 2 + 1 ⋅ (x2 + y2 + z2 ) 2
∂x ∂x
m−2 2 m− 2
−1
m− 2
= m x ⋅ ⋅ (x + y2 + z2 ) 2 ⋅ 2x + (x2 + y2 + z2 ) 2
2
m−4 m−2
= m (m − 2)x2 (x2 + y2 + z2 ) 2 + (x2 + y2 + z2 ) 2
m−4
= m (x2 + y2 + z2 ) 2 (m − 2)x2 + ( x2 + y2 + z2 ) …(3)
Similarly,
m− 4
Vyy = m (x2 + y2 + z2 ) 2 ( m − 2) y2 + (x2 + y2 + z2 )
…(4)
( )
m− 4
Vzz = m (x2 + y2 + z2 ) 2 (m − 2 ) z2 + (x2 + y2 + z2 )
2
…(5)
Adding (3), (4) and (5),
m−4
Vxx + Vyy + Vzz = m (x2 + y2 + z2 ) ⋅ 2 ((m − 2 + 3)(x2 + y2 + z2 ))
m− 4 m− 2
+1
= m(m + 1)(x2 + y2 + z2 ) 2 = m(m + 1)(x2 + y2 + z2 )2 = m(m + 1) rm
–x2
1 4a2t ∂v ∂2 v
Example 3: If v = e , prove that = a2 2
t ∂t ∂x
∂v ∂ − 21 4−ax2t 1 − 23 4−ax2t − 21 4−ax2t 1
2 2 2
= t e = − t e + t e ⋅ − 2 − 2
x2
Solution:
∂t ∂t 2 4 a t
1 − 3 − x2 −
5 − x2
x2
= − t 2 e 4 a2t + t 2 e 4 a2t 2 …(1)
2 4a
3
∂v − 12 4−ax2t 2x t 2 −x
−
x − 3 −x
2 2 2
= t e − 2 = − 2 t 2 e 4 a2t = − 2 x e 4 a2t …(2)
∂x 4a t 2a 2a
266 Engineering Mathematics through Applications
3
−
∂2 v −t 2 4−ax2t − x2
−2x
2
= + 4a2t ⋅
∂x 2
2 2
a e x e
4 a2 t
1 1 − 3 −x22 − 5 x2 −x22
= − 2 t 2 e 4a t + t 2 4a2 e 4a t …(3)
a2
∂2 v 1 ∂v ∂v ∂2 v
Using (1), = 2 ⇒ = a2 2
∂x 2
a ∂t ∂t ∂x
Example 4: Given u = er cosθθ cos(r sinθθ), v = er cosθθ sin(r sinθ
θ), prove that
∂u 1 ∂v ∂v 1 ∂u
= and =− [NIT Kurukshetra, 2009]
∂ r r ∂θ ∂r r ∂θ
Solution:
∂u = er cos θ cos θ {cos(r sin θ)} + er cos θ {− sin(r sin θ) (sin θ)}
∂r θ constant
= er cosθ [cos θ{cos θ(rsin θ)}– sin θ{sin(rsin θ)}] …(1)
∂v = er cos θ (−r sin θ) {sin(r sin θ)} + er cos θ {cos(r sin θ) r cos θ}
And ∂θ
r constant
= ercos θ[–r sin θ{sin(r sin θ)} + r cos θ{cosθ(rsinθ)}]
= rercosθ[– sinθ{sin(rsinθ)} + cosθ{cos(rsinθ)}] …(2)
Using (1),
∂u 1 ∂u
=
∂r r ∂θ
For part (ii), readers may try themselves.
∂2 u ∂2 u
Example 5: Prove that + = 0, if *[NIT Kurukshetra, 2008]
∂x2 ∂y2
2xy
*(i) u = (tan–1a) [log(x2 + y2)] + b tan–1 (y/x), (ii) u = tan–1 2 .
x – y2
Solution: (i) From given, we get
∂u y
= ( tan−1 a ) 2
1 1
2x + b 2 x2 − 2
∂x (x + y )
2
x +y 2 x
∂u 2x
= ( tan−1 a ) ⋅ 2
by
− 2
∂ (x + y ) ( x + y )
or x 2 2
Further,
(x2 + y2 )2 − 2x(2x) (x2 + y2 )0 − y2x
∂2 u
2 =(
tan−1 a) ⋅ −b
∂x
( x2
+ y2
)2
( 2x2 + y2 )
2
Partial Derivatives and their Applications 267
2(y2 − x2 )
= ( tan−1 a ) 2
2bxy
2 2+ …(1)
( x + y ) ( x2
+ y2 )2
∂u 2y x2 1
Again, = ( tan−1 a ) 2 +b 2 .
∂y (x + y )
2
x + y2 x
2y
= ( tan−1 a ) 2
bx
2 +
(x + y ) ( x2
+ y2 )
∂2 u (x2 + y2 )2 − 2y ⋅ 2y b ⋅ x(−1)(2y)
= ( tan −1
a ) +
∂y2 (x2 + y2 )2 ( x2 + y2 ) …(2)
2
2(x2 − y2 ) 2bxy
= ( tan −1 a ) 2 2 2
− 2
(x + y ) (x + y )
2 2
∂2u ∂2u
Adding (1) and (2), we get + =0
∂x2 ∂y2
(ii) Let }
x = r cos θ
y = r sin θ
so that x2 + y2 = r2 and
y
θ = tan−1
x
2xy −1 2 sin θ cos θ −1 sin 2 θ
u = tan−1 2 2 = tan 2 = tan = 2θ
cos 2 θ
…(4)
x − y cos2
θ − sin θ
∂u ∂θ ∂ y 1 y −2y
Therefore =2 = 2 tan−1 = 2 2 − 2 = 2 …(5)
∂x ∂x ∂x x y x x +y
2
1 +
x
∂2u −1 4xy
2 = −2y ⋅ 2x =
and ∂x ( x2 + y2 )2
( x2 + y2 )2 …(6)
∂2u −4xy
2 =
Similarly ∂y ( x + y2 )2
2 …(7)
∂2u ∂2u
On adding (6) and (7), + = 0.
∂ x2 ∂ y 2
Hence the result.
2
∂ ∂ ∂
– 3xyz], show that ∂x + ∂y + ∂ z- u = −9 (x + y + z)
-2
Example 6: If u = log[x3 + y3 + z3
[KUK, 2006; UP Tech, 2006; PTU, 2008]
Solution: Given u = log (x3 + y3 + z- 3 – 3xyz). …(1)
∂ ∂ ∂
We need to prove that the operator ∂x + ∂y + ∂z if applied twice on the function
u(x, y, z) results in –9(x + y + z)–2.
268 Engineering Mathematics through Applications
∂u ∂u ∂u
Therefore first find, ∂x , ∂y , ∂z
∂u ∂
= log (x3 + y3 + z3 − 3xyz)
∂x ∂x
(i.e. partial derivative of u with respect to x keeping y and z– constant)
∂u 1 ∂ 3
= (x + y3 + z3 − 3xyz)
∂x x3 + y3 + z3 − 3xyz ∂x
∂u = 3x2 − 3yz
⇒ ∂x y , z x3 + y3 + z3 − 3xyz …(2)
Similarly we obtain
∂u 3y2 − 3xz
∂y = 3 …(3)
x,z x + y3 + z3 − 3xyz
∂u = 3z2 − 3xy
and ∂z x, y x3 + y3 + z3 − 3xyz …(4)
On adding (2), (3) and (4), we get
∂u ∂u ∂u 3 ( x + y + z − xy − yz − zx )
2 2 2
+ +
∂x ∂y ∂z =
x3 + y3 + z3 − 3xyz
∂u ∂u ∂u 3 ( x2 + y2 + z2 − xy − yz − zx )
∂x + ∂y + ∂z =
or ( x + y + z) ( x2 + y2 + z2 − xy − yz − zx )
∂u ∂u ∂u 3
⇒ ∂x + ∂y + ∂z = x + y + z
( ) …(5)
or U = 3(x + y+ z)–1
Likewise, obtain partial derivatives of expression (5) with respect to x, y, z respectively
∂ ∂
i.e. U= 3(x + y + z)−1
∂x ∂x
∂U
or = − 3(x + y + z)−2 …(6)
∂x
Similarly
∂U
= −3(x + y + z)−2 …(7)
∂y
∂U
and = − 3(x + y + z)−2 …(8)
∂z
Partial Derivatives and their Applications 269
Adding (6), (7), and (8), we get
∂U ∂U ∂U
+ + = −9(x + y + z)−2
∂x ∂y ∂z
∂ ∂ ∂
∂x + ∂y + ∂z U = −9(x + y + z)
−2
or
2
∂ ∂ ∂ −2
or ∂x + ∂y + ∂z u = −9(x + y + z)
Hence the result.
∂2z
Example 7: If xxyyzz = c, show that at x = y = z, = − (x logex)–1 [NIT Jalandhar, 2004]
∂ x∂ y
Solution: Rewrite xxyyzz = c as zz = c/xxyy …(1)
Taking logs,
z log z = log c – x log x – y log y …(2)
Differentiating partially both sides with respect to x
∂z log z + z 1 ∂z = −1 ⋅ log x − x 1
∂x z ∂x x
∂z
i.e.
∂x
(1 + log z) = − (1 + log x ) …(3)
Differentiating both sides of (3) partially with respect to y
∂2 z
∂x∂y
(1 + log z) + ∂∂xz 1z ∂∂yz = 0
∂2 z 1 ∂z ∂z
∂ x∂ y (
log e z ) = − ⋅
z ∂x ∂y
1 ∂z ∂z
∂z2 z ∂x ∂y
=− …(4)
∂ x∂ y log e z
At x = y = z,
1
∂2 z ⋅1⋅1 ∂z ∂z
=− x = −(x log ex)−1 , since ∂x = 1 = ∂y at x = y = z
∂ x∂ y log e x
∂3 u ∂3 u
Example 8: If u = xy, show that =
∂ x2 ∂ y ∂ x ∂ y ∂ x
∂u Using d ax = ax log a, where a is constant
Solution: Given u = xy gives ∂ = xy ⋅ loge x dx
y
∂2 u ∂ ∂u 1
= = yxy −1 ⋅ loge x + xy ⋅ = xy −1(1 + y log x)
∂x∂y ∂x ∂y x
270 Engineering Mathematics through Applications
∂ ∂2 u ∂ y −1
= x ⋅ (1 + y log x)
∂x ∂x∂y ∂x
and …(1)
∂u ∂2u
Again = xy −1 and = 1 ⋅ xy −1 + y(xy −1 log x) = xy −1(1 + y log x) …(2)
∂x ∂y∂x
Using (2),
∂3 u ∂ ∂2 u ∂ y −1
= = x (1 + y log x)
∂x∂y∂x ∂x ∂y∂x ∂x …(3)
From (1) and (3) follows the result.
∂2v ∂2v ∂2v
Example 9: If v = log (x2 + y2 + z2), prove that (x2 + y2 + z2 ) 2 + 2 + 2 = 2
∂x ∂y ∂z
Solution: Given v = log (x2 + y2 + z2)
∂v = 1
2x
⇒ ∂x y , z (x2 + y2 + z2 ) …(1)
∂2 v ∂ ∂v (x2 + y2 + z2 ) ⋅ 1 − x ⋅ 2x (y2 + z2 − x2 )
And 2 = ∂x ∂x = 2 = 2 2 2 2 …(2)
∂x (x + y + z )
2 2 2 2
(x + y + z )
2
∂2 v (x2 + z2 − y2 )
Similarly = 2 (x2 + y2 + z2 )2 …(3)
∂y2
∂2v (x2 + y2 − z2 )
and 2 = 2 2 2 2 …(4)
∂z (x + y + z )
2
Adding (2), (3) and (4), we get
∂2v ∂2v ∂2v (x2 + y2 − x2 ) + (x2 + z2 − y2 ) + (x2 + y2 − z2 )
∂x2 + ∂y2 + ∂z2 = 2 (x2 + y2 + z2 )2
∂2 v ∂2 v ∂2 v 2
⇒ ∂x2 + ∂y2 + ∂z2 = (x2 + y2 + z2 )
∂2v ∂2v ∂2v
or (x2 + y2 + z2 ) 2 + 2 + 2 = 2
∂x ∂y ∂z
Hence the result.
x2 y2 z2
Example 10: If + 2 + 2 = 1, prove that
a +u b +u c +u
2
2 2 2
∂u + ∂u + ∂u = 2 x ∂u + y ∂u + z ∂u
∂x ∂y ∂z ∂x ∂y ∂z [UP Tech, 2003]
Partial Derivatives and their Applications 271
x2 y2 z2
Solution: Differentiating + + 2 = 1 with respect to x
a +u b +u c +u
2 2
2x x2 ∂u y2 ∂u z2 ∂u
we get − 2 2 ∂x − 2 ∂x − 2 =0
a + u ( a + u)
2
(b + u)
2
(c + u)2 ∂x
2x x2 y2 z2 ∂u
= + +
i.e. a2 + u (a2 + u)2 (b2 + u)2 (c2 + u)2 ∂x
2x ∂u , x2 y2 z2
=K K= 2 + +
(a + u)2 (b2 + u)2 (c2 + u)2
i.e. where …(1)
a2 + u ∂x
2y ∂u
Similarly =K …(2)
b +u
2 ∂y
2z ∂u
and =K …(3)
c2 + u ∂z
On squaring (1), (2), (3) and adding, we get
x2 z2 2
∂u ∂u
2 2
y2 2 ∂u
+ + = + +
∂x ∂y ∂z
4 2 K
(a + u)
2
(b2 + u)2 (c2 + u)2
∂u 2 ∂u 2 ∂u
2
4 ∂u
2
∂u
2
∂u
2
or 4K = K 2 + + , i.e. = + + …(4)
∂x ∂y ∂z K ∂x ∂y ∂z
On multiplying both sides of (1) by 2x, (2) by 2y, (3) by 2z and adding,
x2 y2 z2 ∂u ∂u ∂u
4 2 + 2 + 2 = K ⋅ 2 x +y +z
a + u b + u c + u ∂x ∂y ∂z
Using the given relation and on simplification,
∂u ∂u ∂u 4 ∂u ∂u ∂u
4 = 2K x +y +z or = 2 x +y +z …(5)
∂x ∂y ∂z K ∂x ∂y ∂z
From (4) and (5), we get the desired result
2 2 2
∂u + ∂u + ∂u = 2 x ∂u + y ∂u + z ∂u
∂x ∂y ∂z ∂x ∂y ∂z
Example 11: Find the value of n so that the equation V = rn(3cos2θ – 1) satisfies the relation
∂ 2 ∂V 1 ∂ ∂V
r + sinθ = 0.
∂r ∂r sinθ ∂θ ∂θ
∂V
Solution: Given V = rn (3 cos2θ – 1) so that ∂r = n r (3 cos θ − 1)
n −1 2
θ
272 Engineering Mathematics through Applications
∂V = n +1(3 cos2 θ − 1)
and r2 nr …(1)
∂r
∂ 2 ∂V
Now r = n(n + 1) rn(3 cos2 θ − 1) = n(n + 1)V …(2)
∂r ∂r
∂V = n − ∂V
Further, r ( 6 cos θ sin θ) so that sin θ = − 6rn(cos θ sin2 θ) …(3)
∂θ r ∂θ
∂ ∂V
and sin θ = − 6rn (− sin θ sin2 θ + cos θ ⋅ 2 sin θ cos θ)
∂θ ∂θ
= –6rnsinθ(2 cos2θ – sin2θ) …(4)
1 ∂ ∂V
Implying sin θ ∂θ sin θ ∂θ = − 6r (2 cos θ − sin θ)
n 2 2
= –6rn(2cos2θ + cos2θ – 1)
= –6rn(3cos2θ –1) = – 6V, (using given relation) …(5)
On adding expressions (2) and (5), we get
n(n + 1)V – 6V = 0 or [n(n + 1) – 6]V = 0
which implies either (n2 + n – 6) = 0 or V = 0 (but V ≠ 0)
∴ n2 + 3n + 2n – 6 = 0 or (n + 3) (n - 2) = 0 or n = 2, –3.
∂u ∂2 u
Example 12: The equation = µ 2 refers to the conduction of heat along a bar without
∂t ∂x
–gx
radiation, show that if u = Ae sin(nt – gx), where A, g, n are positive constants, then
n
g=
2µ
Solution: Given, u = A e–gx sin (nt – gx) …(1)
∂u
so that = A e− gx cos(nt − gx) ⋅ n = n A e− gx cos(nt − gx) …(2)
∂t
∂u
and = A e− gx (− g)sin(nt − gx) + e− gx cos(nt − gx)(− g)
∂x
= –gA e–gx [sin (nt – gx) + cos (nt – gx)] …(3)
∂2u
implying = g2 A e− gx sin(nt − gx) + cos(nt − gx) + g2A e–gx [cos(nt – gx) – sin(nt – gx)]
∂x2
= 2g2A e–gx [cos(nt – gx)] …(4)
∂u ∂2 u
Now =µ 2 ⇒ n Ae− gx cos(nt − gx) = 2µg2 A e− gx cos(nt − gx)
∂t ∂x
Partial Derivatives and their Applications 273
or A cos (nt – gx) e–gx [n – 2µg2 ] = 0
Implies either A cos (nt – gx) e –gx = 0 or (n – 2µg2) = 0
2µ 2µ
But A cos(nt – gx) ≠ 0 ∴ g2 = or g=
n n
ASSIGNMENT 1
1. Show that zxy = zyx if z = x3 + y3 – 3axy.
y x ∂2u x2 − y2
2. If u = x2 tan−1 − y2 tan−1 , show that = 2
x y ∂x∂y x + y2
∂2 z 2 ∂ z
2
3. If z = f(x + ct) + φ(x – ct), prove that = c
∂t2 ∂x2
1 ∂2 v ∂2 v ∂2 v
4. If v = (x2 + y2 + z2 ) 2 , prove that ∂x2 + ∂y2 + ∂z2 = 0
−
[Osmania, 2003; VTU, 2005]
∂3 u
5. If u = exyz, find the value of ∂x∂y∂z . [Osmania, 2003; Rajasthan, 2005]
+ , show that x ∂u + y ∂u + z ∂u = 0
y z
6. If u =
z x ∂x ∂y ∂z
2
∂z ∂z ∂z ∂z
7. If z(–x + y) = x2 + y2, show that ∂x − ∂y = 4 1 − ∂x − ∂y [VTU, 2003]
∂V ∂V ∂V
8. If V = log (tan x + tan y + tan z), find the value of sin 2x + sin 2y + sin 2z
∂x ∂y ∂z
(x2 + y2 )
, verify ∂ u = ∂ u .
2 2
9. If u = log [Hint: u = log(x2 + y2) – log xy]
xy ∂y∂x ∂x∂y
∂ ∂u ∂ 2 ∂u
(1 − x2 ) +
1
10. If u = (1 − 2xy + y2 )− 2 , prove that =0
∂x ∂y ∂y
y [MDU, 2006]
∂x
x y ∂u ∂u
11. If u = sin −1 + tan −1 , show that x +y =0
y x ∂x ∂y
12. Verify that fxy = fyx when f(x, y) = (log x )tan–1 (x2 + y2)
13. If Z = log (ex + ey), show that rt – s2 = 0, where r = Zxx, s = Zxy, t = Zyy.
−
r2 1 ∂ 2 ∂θ ∂θ
14. If θ = tne 4t , what value of n will make r = ? [KUK, 2006; UP Tech, 2006]
r2 ∂ r ∂ r ∂ t
274 Engineering Mathematics through Applications
4.3 WHICH VARIABLE TO BE KEPT CONSTANT
} we get r = (x2 + y2 )2
1
For x = r cos θ
…(i), …(ii)
y = r sin θ y
θ = tan −1
x
The above relations clearly indicate that x is a function of (r, θ) in x = r cos θ and x is a
function of (r, y) in r2 = x2 + y2 or x2 = r2 – y2.
∂x
We find either from x = r cosθ (keeping θ constant) or from x2 = r2 – y2 (keeping y
∂r
constant).
∂x = cos θ and 2x ∂x = 2r ⇒ ∂x = r = r = 1
∴ ∂r θ ∂r y ∂r y x r cos θ cos θ
and there is no reason for them to be equal since we have to be specific while finding the
partial derivative that which variable is to be kept constant.
∂r ∂r
Likewise, we find
∂y
either from (i), r = y cosec θ, i.e. ∂y = cosec θ
θ
∂r y r sin θ
Or from (ii), r2 = (x2 + y2), i.e. ∂y = r = r = sin θ
x
Thus, when no indication is given regarding the variable to be kept constant, then by
∂ ∂ ∂ ∂
convention, we take means and means Likewise, ∂ means ∂
∂x ∂x y ∂y ∂y x ∂r ∂r θ
∂ ∂
and means .
∂θ ∂θ r
Example 13: If u = lx + my, v = mx – ly, show that
∂u ∂x = l2 , ∂y ∂v = (l2 + m2 )
∂x y ∂u v l2 + m2 ∂v u ∂y x l2
Solution: Given relations imply,
lu = l2x + lmy …(i)
mv = m2x – mly …(ii)
(lu + mv)
On adding the two (lu + mv) = (l2 + m2)x ⇒ x= …(1)
(l2 + m2 )
Again given implies
mu = mlx + m2y …(iii)
lv = lmx – l2y …(iv)
(mu − lv)
On subtracting (iv) from (iii), (mu – lv) = (l2 + m2)y ⇒ y= …(2)
(l2 + m2 )
Partial Derivatives and their Applications 275
Also from given relations,
∂ u = l ∂ v = −l
∂x y , ∂y …(3)
x
From (1) and (2),
∂x = l ∂y l
∂u v (l2 + m2 ) , ∂v = − (l2 + m2 ) …(4)
u
From above, we can deduce the desired results.
x2 = au + bv ∂u ∂x 1 ∂v ∂y
Example 14: If , prove that ∂x ∂u = 2 = ∂y ∂v
y = au − bv
2
y v x u
Solution: From the given relation,
∂x = ⋅ + ⇒ ∂x = a
2x a 1 0 …(1)
∂u v ∂u v 2x
∂y ∂y b
and 2y = 0 − b ⋅ 1 ⇒ = − …(2)
∂v u ∂v u 2y
1 2
On adding the given equations, u = (x + y2 ) …(3)
2a
1 2
On taking difference of given equations, v = (x − y2 ) …(4)
2b
Partial derivative of (3) with respect to x, ∂ u = 1 2x = x …(5)
∂ x y 2a a
∂v 1 y
Partial derivative of (4) with respect to y, ∂y = 2b ( − 2y) = − b …(6)
x
Multiplying (1) and (5), (2) and (6), we get the desired results.
} ∂2r ∂2r 1 ∂r ∂r
2 2
x = r cosθ
Example 15: If , prove that (i) + = + ∂y
y = r sinθ ∂ x2 ∂ y2 r ∂x
∂2 θ ∂2 θ (x ≠ 0, y ≠ 0)
(ii) + =0 [ KUK, 2004, 2008; NIT Kurukshetra, 2006]
∂x2 ∂y2
Solution:
x = r cos θ
y = r sin θ } ⇒
r2 = x2 + y2
y
θ = tan−1
x
…(A)
…(B)
∂r ∂r x ∂r x
From (A), 2r = 2x or = or = …(1)
∂x ∂x r ∂x (x2 + y2 )1 2
276 Engineering Mathematics through Applications
Differentiating (1) again partially with respect to x, we get
x2
(x2 + y2 ) −
∂r (x + y2 ) 2 (x2 + y2 ) − x2
1
2 2
y2
= = 2 3 2 =
…(2)
∂x2 (x + y )
2 2
(x + y )
2
(x + y2 )3
2 2
∂y y
Similarly, = …(3)
∂r r
∂2 y x2
and 2 = …(4)
∂r (x2 + y2 )3 2
Adding (2) and (4), we get
∂2 r ∂2 r y2 x2
2 + 2 = +
∂x ∂y (x + y2 )3
2 2
(x + y2 )3
2 2
1 y2 x2
= (x2 + y2 ) + (x2 + y2 )
(x + y2 )1
2 2
1 ∂r ∂r
2 2
= +
r 2 ∂y ∂x , Using (1) and (3)
Hence the result.
Part (ii):
From the given relation, θ = tan −1 ,
y
x
∂θ = 1 ∂ y x2 −y y
∂x y 2 ∂x
x = x2 + y2 ⋅ x2 = − x2 + y2
∴ y …(5)
1+
x
Again differentiating it partially with respect to ‘x’,
∂2θ 1 2xy
= −y(−1) 2 2 2 2x
= 2 …(6)
∂x2
(x + y ) (x + y2 )2
∂θ 1 ∂ y x2 1 x
∂y = 2 ∂y
x
= 2
+ 2 ⋅x = 2
+ y2
1 +
Similarly, x y x y x …(7)
x
∂2θ −x −2xy
and = ⋅ 2y = 2 …(8)
∂y2 (x2 + y)2 (x + y2 )2
∂2θ ∂2θ 2xy −2xy
Adding the two, + 2 = 2 + 2 = 0 , provided x ≠ 0, y ≠ 0.
∂x2
∂y (x + y ) (x + y2 )2
2 2
Partial Derivatives and their Applications 277
Example 16: If u = f(r), where r2 = (x2 + y2 + z2), show that
∂2 u ∂2 u ∂2 u 2
+ + = f''(r ) + f'(r) [UP Tech, 2005; Rajasthan, 2006]
∂ x2 ∂ y2 ∂ z2 r
Solution: Here u = f(r), where r2 = (x2 + y2 + z2) …(1)
∂u ∂r
Therefore = f'(r) …(2)
∂x ∂x
∂2u ∂ ∂r = f"(r) ∂r ⋅ ∂r + f'(r) ⋅ ∂2 r
and = f'(r)
∂x2 ∂x ∂x ∂x ∂x ∂ x2
2
∂r ∂2 r
= f"(r) + f'(r) 2 …(3)
∂x ∂x
2
∂2 u ∂r ∂2 r
2 = f"(r ) ∂y + f'(r ) ⋅
Similarly …(4)
∂y ∂y2
2
∂2u ∂r ∂2r
and 2 = f"(r ) ∂z + f'(r ) …(5)
∂z ∂z2
Adding (3), (4) and (5), we get
∂2u ∂2u ∂2u ∂r 2 ∂r 2 ∂r 2 ∂2r ∂2r ∂2r
+ + = f"(r ) ∂x + ∂y + ∂z + f'(r) ∂x2 + ∂y2 + ∂z2 …(6)
∂x2 ∂y2 ∂z2
∂r ∂ 2 1 1 x
Now = (x + y2 + z2 ) 2 = ⋅ 2x =
∂x ∂x 2 12
2(x + y + z )
2 2 r
1 1
(x2 + y2 + z2 ) 2 ⋅1− x 1 ⋅ 2x
∂r
2 2(x + y2 + z2 )
2 2
Further =
∂x2 (x + y + z2 )
2 2
(x2 + y2 + z2 ) − x2 r2 − x 2
= 3 =
(x2 + y2 + z2 ) 2 r3
∂r x ∂2r r2 − x2
Precisely = , = ; …(7)
∂x r ∂x2 r3
∂r y ∂2r r2 − y2
= , =
∂y r ∂y2 r3 ; …(8)
∂r z ∂2 r r2 − z2
= , = …(9)
∂z r ∂z2 r3
278 Engineering Mathematics through Applications
On using results given by (7), (8), (9) into equation (6), we get
∂2u ∂2u ∂2u x 2 y 2 z 2 r2 − x2 + r2 − y2 + r2 − z2
+ +
∂x2 ∂y2 ∂z2 = f"(r ) r + r + r + f'(r)
r3
x2 + y2 + z2 3r2 − r2 2
= f"(r) + f'(r) = f"(r) + f'(r)
r 2
r 3 r
ASSIGNMENT 2
2 2
r
1. If x = r cos θ, y = r sin θ, show that (i) ∂ = ∂ (ii) 1 ∂ = r ∂θ , (iii) ∂ + ∂ = 1 .
r x x r
∂x ∂r r ∂θ ∂x ∂x ∂y
x = r cos θ
}
2. If u = f (r) and y = r sin θ , prove that
∂2 u ∂2 u
+
∂x2 ∂y2
1
= f"(r) + f'(r ) .
r
[Burdwan, 2003]
cos θ sin θ ∂y
3. If x = ,y = , show that ∂x ∂u + ∂u = 1 .
u u ∂u θ ∂x y ∂u θ ∂y x
4.4 HOMOGENEOUS FUNCTIONS AND EULERS THEOREM
Homogeneous Function
An expression of the form (a0xn + a1xn – 1y + a2xn – 2y2 + …+ anyn) in which all the terms are of
degree n, is called a homogeneous function of degree n.
The above expression mathematically may be rewritten as
xn[a0 + a1(y/x) + a2(y/x)2 + … + an(y/x)n] or more precisely xnφ(y/x),
where φ(y/x) is a polynomial of degree n in (y/x).
Thus, any function f(x, y) which is expressible as either xn φ(y/x) or ynψ(x/y) is called a
homogeneous function of degree n in x and y.
E.g. (i) x3 tan (y/x) is a homogeneous function of degree 3 in x and y.
1 y 1 3
x 1 +
3
1 1
x 3+y 3 x
(ii) = = x1 3 −1 2 φ (y / x) = x−1 6 φ(y / x) is a homogeneous function of
1 y 2
1 1 1
x 2+y 2
x 2 1 +
x
degree –1/6 in x and y.
In general, a function f(x1, x2, x3, …) is said to be a homogeneous function of degree n in
x x x
(x1, x2, … , xn) if it is expressible in the form as x1n φ 2 , 3 , …, n .
x1 x1 x1
Euler’s Theorem
If u be homogeneous function of degree n in x, y and has continuous first derivatives then
∂u ∂u
x +y = nu. [KUK, 2004]
∂x ∂y
Partial Derivatives and their Applications 279
y
Proof: u is a homogeneous function of degree n in x, y, i.e. u = xn φ .
x
∂u y y −1 y y
∴ = n xn −1 φ + xn φ′ y ⋅ 2 = n xn −1 φ − xn − 2 y φ′ …(1)
∂x x x x x x
∂u y 1 y
and = xn φ' = xn −1 φ' …(2)
∂y x x x
∂u ∂u y
From (1) and (2), x +y = n xn φ = nu . Hence the result.
∂x ∂y x
In general, if u be a homogeneous function of degree n in x1, x2, x3, … xn, then
∂u ∂u ∂u ∂u
x1 + x2 + x3 + … + xn = nu.
∂x1 ∂x2 ∂x3 ∂xn
∂u ∂u
Example 17: If u = sin–1x/y + tan–1y/x, prove that x +y = 0.
∂x ∂y
Solution: We have u = sin–1x/y + tan–1y/x
1
= x° sin−1 −1
+ tan (y / x) comparable to x f (y / x),
n
( y / x)
where n = 0
∴ u is a homogeneous function of order 0.
∂u ∂u
Hence Euler’s Theorem, x +y = nu = 0, (as n = 0).
∂x ∂y
(x2 + y2 ) ∂ψ ∂ψ
Example 18: If ψ( x, y) = log (x + y) , show that x +y = 1.
∂x ∂y
(x2 + y2 )
x2 1 + ( y / x)2
Solution: We have ψ(x, y) = log (x + y) = log
( ) (
1 + ( y / x )2
)
( ) = log x 1 + ( y / x )
( )
x + ( y x)
1 /
1 + ( y / x )2
ψ (x , y )
= x comparable to xn φ (y / x), where n = 1.
( )
So that e
1 + ( y / x )
Thus eψ(x, y) is a homogeneous function of degree 1 in x and y.
By Euler’s Theorem,
∂ ψ( x , y ) ∂
x⋅ (e ) + y (eψ(x, y) ) = 1 ⋅ eψ(x , y)
∂x ∂y
∂ψ ∂ψ ∂ψ ∂ψ
⇒ x eψ(x , y) + y eψ(x , y ) = eψ(x, y) or x +y =1
∂x ∂y ∂x ∂y
280 Engineering Mathematics through Applications
(x2 y2 ) ∂u ∂u
Example 19: If sin u = , show that x +y = 3tan u
( x + y) ∂x ∂y
y y
2
2
x4
x x
Solution: sin u = = x3 = x3 φ y is a homogeneous function of degree
y
x 1 +
y x
x 1 + x
3 in x and y.
Therefore by Euler’s Theorem,
∂ ∂
x (sin u) + y (sin u) = 3 ⋅ sin u
∂x ∂y
∂u ∂u
⇒ x cos u + y cos u = 3 sin u
∂x ∂y
∂u ∂u
⇒ x +y = 3 ⋅ tan u
∂x ∂y
( x + y)
Example 20: If cos−1 = u, then show that x ∂u + y ∂u = − 1 cot u [VTU, 2004]
x+ y ∂x ∂y 2
(x + y) x+y
Solution: Here u = cos−1 or = cos u = z
x+ y x+ y
1
2
2
y 1 + y
x 1 + x
where z=
x+y
= x = x1 2 = x 2 φ y
1
x+ y y 1 2 x
y 2
1
x 1 + 1 + x
x
Thus z is a homogeneous function of degree 1 in x and y.
2
By Euler’s Theorem,
∂z ∂z ∂u ∂u 1 1
x +y = nz or x(− sin u) + y(− sin u) = z = cos u
∂x ∂y ∂x ∂y 2 2
∂u ∂u −1
⇒ x +y = cot u
∂x ∂y 2
1
1 2 + y1 2 2
−1 x
Example 21: If u = cosec 1 1 ,
prove that
x 3 + y 3
∂2 u ∂2 u 2 ∂ u
2 tan u 13 tan2 u
x2 2 + 2 xy ∂x ∂y + y 2 = 12 12
+
12 [MDU, 2006]
∂x ∂y
Partial Derivatives and their Applications 281
1
x1 2 + y1 2 2
1 = z
Solution: Let 1 so that cosec u = z
x 3 +y 3
1 1
1 1 2
0
1 2
x 2 1 + y
1 2 y y 2
+ x
12
( ) 1
1 2
x + y2 x 1 1
−
2
x
Now z= 1 = = x 2 3
1
3 1 y 2
1 1 0
x + y3 y + y2
x2 1 + x x
x
1
y y
= x12 Function with all terms of the form = xn φ
x x
1
Clearly a homogeneous function of degree n = in x and y.
12
Therefore by Euler’s Theorem,
∂ ∂
x (z) + y (z) = nz
∂x ∂y
∂ ∂ 1
x (cos ec u) + y (cos ec u) = cosec u
∂x ∂y 12
∂u ∂u 1
x ⋅ − cos ec u ⋅ cot u + y ⋅ − cos ec u ⋅ cot u = cosec u
∂x ∂y 12
∂u ∂u −1
or x ∂x + y ∂y = 12 tan u …(1)
Now on differentiating (1) partially with respect to x and with respect to y respectively,
we get
∂ ∂u ∂u 1 ∂
+y =−
∂x ∂x
x (tan u)
∂y 12 ∂x
∂2u ∂u ∂2u 1 2 ∂u
2 + 1 ⋅ ∂x + y ∂x∂y = − 12 sec u ∂x
or x …(2)
∂x
∂ ∂u ∂u 1 ∂
and +y =−
∂y ∂x
x (tan u)
∂y 12 ∂y
∂2u ∂2u ∂u 1 ∂u
or x + y 2 + 1⋅ = − sec2 u …(3)
∂x∂y ∂y ∂y 12 ∂y
Multiply (2) by x and (3) by y and then add the two,
2 ∂2u ∂2u 2 ∂ u
2
∂u ∂u 1 2 ∂u ∂u
x ∂x2 + 2xy ∂x∂y + y ∂y2 + x ∂x + y ∂y = − 12 sec u x ∂x + y ∂y
282 Engineering Mathematics through Applications
2 ∂2 u ∂2 u 2 ∂ u
2
1 ∂u ∂u
x ∂x2 + 2xy ∂x∂y + y ∂y2 = − 12 sec u − 1 x ∂x + y ∂y
or 2
= − ( tan2 u + 1) − 1 −
1 tan u
12 12 On using (1)
∂2u ∂2u ∂2u tan u 13 tan2 u
Hence the result x2 2 + 2xy ∂x∂y + y2 2 = 12 12 + 12
∂x ∂y
∂2 z ∂2 z ∂2 z
Example 22: If z = x φ (y/x) + ψ (y/x), prove that x2 + 2xy + y2 2 = 0
∂x 2 ∂x ∂y ∂y
[UP Tech, 2006]
Solution: Let us write z = xφ (y/x) + ψ(y/x) = z1 + z2,
z 1 = x φ (y/x) = x1 φ (y/x) is a homogeneous function of degree 1 in x and y
z2 = ψ (y/x) = x0 ψ (y/x) is a homogeneous function of degree 0 in x and y
∂z1 ∂z ∂z1 ∂z
∴ x + y 1 = n ⋅ z1 ⇒ x + y 1 = 1 ⋅ z1 = z1 …(1)
∂x ∂y ∂x ∂y
∂z2 ∂z ∂z2 ∂z
and x + y 2 = n ⋅ z2 ⇒ x + y 2 = 0 ⋅ z2 = 0 …(2)
∂x ∂y ∂x ∂y
∂2 z1 ∂z1 ∂2 z1 ∂z1
Now differentiating (1) with respect to x, x + + y =
∂x2 ∂x ∂ x∂ y ∂ x
∂2 z1 ∂2 z1
⇒ x + y =0 …(3)
∂x 2 ∂ x∂ y
∂2 z1 ∂z1 ∂2 z ∂z
Differentiating (1) with respect to y, x ⋅ + + y 21 = 1
∂ y∂ x ∂ y ∂y ∂y
∂2 z1 ∂2 z
⇒ x⋅ + y 21 = 0 …(4)
∂y ∂x ∂y
Now multiply (3) by x and (4) by y and, add the two
∂2 z ∂2 z1 ∂2 z1 ∂2 z1
x x 21 + y + y x + y =0
∂x ∂x∂y
∂ y∂ x ∂y2
2 ∂2 z1 2 ∂ z1
2
∂2 z1
⇒ x ∂ x2 + y ∂ y2 + 2 xy ∂ x∂ y = 0 …(5)
∂2 z2 ∂z2 ∂2 z2
Similarly on differentiating (2) with respect to x, x + + y =0 …(6)
∂ x2 ∂ x ∂x∂y
Partial Derivatives and their Applications 283
∂2 z ∂2 z ∂ z2
Differentiating (2) with respect to y, x + y 22 + =0 …(7)
∂y ∂x ∂y ∂y
Multiply (6) by x and (7) by y and then add the two, we get
2 ∂2 z2 2 ∂ z2
2
∂2 z2 ∂z2 ∂z2
x ∂x2 + y ∂y2 + 2xy ∂x∂y + x ∂x + y ∂y = 0
2 ∂2 z2 2 ∂ z2
2
∂2 z2
⇒ x
∂x2 + y + 2xy = 0, using (2) …(8)
∂y2 ∂x∂y
Now adding (5) and (8),
∂2 z ∂2 z ∂2 z ∂2 z ∂2 z1 ∂2 z2
x2 21 + 22 + y2 21 + 22 + 2xy + =0
∂x ∂x ∂y ∂y ∂x∂y ∂x∂y
∂2 ∂2 ∂2
or x2 2 (z1
+ z2 ) + y2 2 (z1 + z2 ) + 2xy (z + z ) = 0
∂x ∂y ∂x∂y 1 2
∂2 z ∂2 z ∂2 z
Hence the desired result, x2 + y2 2 + 2xy = 0.
∂x 2
∂y ∂xy
Note: As here in this example, z = z1 + z2, is a sum of two homogeneous functions of diffferent degree of
homoginity. Hence apply Euler’s theorem on z1 and z2 separately upto second order and then finally add them
to have the desired result.
x3 + y3 ∂u ∂u
Example 23: If u = tan−1 , then show that x +y = sin 2x
x−y ∂x ∂y
∂2 u ∂2 u
+ y2 2 + 2xy ∂ u = sin 4u − sin 2u
2
and x2
∂x 2
∂y ∂x ∂y
[NIT Kurukshetra, 2004; VTU, 2005; SVTU, 2007; KUK, 2007-2009; PTU, 2009]
y
3
x3 1 +
(x3 + y3 ) (x3 + y3 ) x
Solution: u = tan−1 ⇒ tan u = = = x2 φ y
x+y (x + y) y x
x 1 +
x
Thus, tan u is a homogeneous function of degree 2 in x and y.
∴ By Euler’s Theorem, x ∂ (tan u) + y ∂ (tan u) = 2 ⋅ tan u
∂x ∂y
∂u ∂u
x sec2 u + y sec2 u = 2 tan u
∂x ∂y
∂u ∂u sin u
x +y =2 × cos2 u = 2 sin u cos u
∂x ∂y cos u
284 Engineering Mathematics through Applications
∂u ∂u
x +y = sin 2u
∂x ∂y …(1)
Differentiating both sides of (1) partially with respect to x,
∂2u ∂u ∂2u ∂u
x 2 + ∂x + y ∂x∂y = 2 cos 2u ∂x
∂x
∂2u ∂2u ∂u
∴ x 2 + y ∂x∂y = (2 cos 2u − 1) ∂x …(2)
∂x
∂2u ∂2u ∂u
Similarly y +x = (2 cos 2u − 1) …(3)
∂y 2 ∂x∂y ∂y
Multiply (1) by x, (2) by y and add the two,
∂2u ∂2u 2 ∂ u
2 ∂u ∂u
x2 2 + 2xy ∂x∂y + y = (2 cos 2u − 1) x
∂
+y
∂y
∂x ∂ y2 x
∂2u ∂2u 2 ∂ u
2
x2 + 2xy + y = (2 cos 2u − 1) [ sin 2u] , Using (1)
∂ x2 ∂x∂y ∂ y2
= [ 2 cos 2u sin 2u − sin 2u = sin 4u − sin 2u]
Example 24: Given Z = xnf1 (y/x) + y–nf2(x/y), prove that
∂2 Z ∂2 Z ∂2 Z ∂Z y ∂Z = n2 Z
x2 + 2xy + y2 2 + x + ∂y
∂x 2 ∂ x∂ y ∂y ∂x
[NIT Kurukshetra, 2005; MDU, 2003, 2006; KUK, 2009]
Solution: Let Z = Z1 + Z2,
Here Z1 = xnf1(y/x), clearly a homogeneous function of degree n in x and y.
Z2 = y–nf2(x/y), clearly a homogeneous function of degree –n in x and y.
∂Z1 ∂Z
Therefore by Euler’s Theorem, x + y 1 = n Z1 …(1)
∂x ∂y
∂Z2 ∂Z
and x + y 2 = −n Z2 …(2)
∂x ∂y
Now differentiating (1) partially with respect to x and y, we have
∂2 Z1 ∂Z1 ∂2 Z1 ∂Z
x + + y =n 1 …(3)
∂x2 ∂x ∂y∂x ∂x
∂2 Z1 ∂Z1 ∂2 Z ∂Z
and x + + y 21 = n 1 …(4)
∂y∂x ∂y ∂y ∂y
Partial Derivatives and their Applications 285
Now multiply (3) by x and (4) by y and, then add the two, we get
∂2 Z1 ∂Z1 ∂Z1 ∂2 Z1 2 ∂Z1 ∂Z ∂Z
x 2 + x ∂x + y ∂y + 2xy ∂y∂x + y = n x 1 + y 1
∂ ∂y
∂x ∂y2 x
∂2 Z1 ∂Z1 ∂Z1 ∂2 Z 2 ∂ Z1
2
⇒ x 2 + x ∂x + y ∂y + 2xy ∂y∂x + y = n nZ1 = n2 Z1 …(5)
∂x ∂y2
Similarly on differentiating (2) partially with respect to x (when y constant) and with
respect to y (when x constant),
∂2 Z2 ∂Z2 ∂2 Z2 ∂Z2
x 2 + ∂x + y ∂y∂x = −n ∂x …(6)
∂x
∂2 Z2 ∂Z2 ∂2 Z ∂Z
and x + + y 22 = −n 2 …(7)
∂ x∂ y ∂ y ∂y ∂y
On multiplying (6) by x and (7) by y and, then adding the two
∂2 Z2 ∂Z2 ∂Z2 ∂2 Z2 2 ∂ Z2
2
∂Z2 ∂Z2
x2 + x + y + 2xy + y 2 = −n x ∂x + y ∂y = −n(−n)Z2 = n Z2
2
…(8)
∂x 2
∂ x ∂ y ∂ x ∂ y ∂y
On adding (5) and (8), we have
∂2 Z ∂2 Z 2 ∂ Z
2 ∂Z ∂Z
x2 + + + x + y = n2 Z
∂y2 ∂x
2 xy y
∂ x2 ∂x∂y ∂y
∂2 u ∂2 u 2 ∂ u
2
2 + 2 xy ∂ x∂ y + y = 2u
Example 25: If u = x2tan–1(y/x) – y2tan–1(x/y), evaluate x2
∂x ∂ y2
[GJU, 2005]
Solution: u = u1 + u2, where u1 = x2tan–1(y/x), u2 = –y2tan–1(x/y) are both homogeneous
function of degree 2 in x and y.
∂u1 ∂u
∴ By Euler’s Theorem, x + y 2 = 2 ⋅ u1 …(1)
∂x ∂y
∂u2 ∂u
and x+ y 2 = 2 ⋅ u2 …(2)
∂x ∂y
Differentiating (1) partially with respect to x and y
∂2u1 ∂u1 ∂2u1 ∂u1
x 2 + ∂x + y ∂x∂y = 2 ∂x …(3)
∂x
∂2u1 ∂u1 ∂2u ∂u
and x + + y 21 = 2 1 …(4)
∂y∂x ∂y ∂y ∂y
Now multiply (3) by x, (4) by y and, add the two
2 ∂2u1 ∂u1 ∂2u1 ∂ u1 2 ∂ u1
2
∂u1 ∂u1
x ∂x2 + x ∂x + 2xy ∂x∂y + y ∂y + y ∂y2 = 2 x ∂x + y ∂y
286 Engineering Mathematics through Applications
2 ∂2 u1 ∂2 u1 2 ∂ u1
2
∂ u1 ∂ u1
⇒ x ∂x2 + 2xy ∂x∂y + y ∂y2 = x ∂x + y ∂y = 2u1 , (using (1))
∂2u2 ∂2u2 2 ∂ u2
2
Similarly x ∂x2 + 2xy ∂x∂y + y ∂y2 = 2u2 , (using (2))
On adding the two, we get the desired result.
Example 26: If u is a homogeneous function of nth degree in x, y, z, prove that
∂u ∂u ∂u
x +y +z = nu [NIT Kurukshetra, 2006]
∂x ∂y ∂z
Solution: u is a homogeneous function of degree n in x, y and z
y z y z
Let u = xn f , or u = xnf (s, t) with = s and =t …(1)
x x x x
Differentiating (1) partially with respect to x
∂u ∂f ∂s ∂f ∂t
= n xn −1 f (s, t) + xn +
∂x ∂s ∂x ∂t ∂x
∂u ∂f y ∂f z
= n xn −1 f (s, t) + xn − 2 + − 2 using (1)
∂x ∂s x ∂t x
∂u ∂f ∂f
⇒ = n ⋅ xn −1 f (s, t) − xn − 2 y +z …(2)
∂x ∂ s ∂t
On multiplying (2) by x throughout, we get
∂u ∂f ∂f
x = n xn f (s, t) − xn −1 y +z …(3)
∂x ∂ s ∂t
Now on differentiating (1) partially with respect to y, we get
∂u ∂f ∂s ∂f ∂t
= xn +
∂y ∂s ∂y ∂t ∂y
∂u ∂f 1 ∂f ∂f
⇒ = xn + (0) = xn −1 …(4)
∂y ∂ s x ∂ t ∂s
On multiplying (4) by y throughout, we get
∂u ∂f
y = xn −1 ⋅ y …(5)
∂y ∂s
Similarly, on differentiating equation (1) with respect to z, we get
∂u ∂f ∂s ∂f ∂t
= xn +
∂z ∂s ∂z ∂t ∂z
Partial Derivatives and their Applications 287
∂f ∂f 1 ∂f
= xn ⋅ (0) + ⋅ = xn −1 …(6)
∂s ∂t x ∂t
On multiplying (6) by z throughout, we get
∂u ∂f
= xn −1 ⋅ z
z …(7)
∂z ∂t
Adding equations (3), (5) and (7), we get
∂u ∂u ∂u
x +y +z = n xn f (s, t) = nu.
∂x ∂y ∂z
Hence the proof.
ASSIGNMENT 3
x+y
1. If u = 2 , then prove that the order of u is –1.
x + y2
2. If u is a homogeneous function of nth degree in x, y, z,
x y z ∂u ∂u ∂u
(a) if u = + + , then x +y +z = 0.
y+z z+x x+ y ∂x ∂y ∂z
x3 + y3 + z3
(b) show that x ux + y uy + z uz = 2 tan u, where u = sin−1 ax + by + cz
xy + yz + zx
3. If u = cos 2 , prove that x ∂u + y ∂u + z ∂u = 0 .
x + y2 + z2 ∂x ∂y ∂z
y ∂2u ∂2u
4. If u = f (x, y) = log(x2 + y2 ) + tan−1 , prove that + = 0.
x ∂ x2 ∂ y 2
∂u ∂u x3 + y3
5. Show that x +y = 2u log u, where log u = .
∂x ∂y 3x + 4y
x + 2y + 3z ∂u ∂u ∂u
6. If u = sin−1 , find the value of x ∂x + y ∂y + z ∂z . [UP Tech, 2004]
x8 + y8 + z8
7. If –z is a homogeneous function of degree n in x and y, show that
∂2 z 2 ∂ z
2
∂2 z
x2 + y + 2 xy = n(n − 1)z. [Kottayam 2005; UP Tech, 2006; VTU, 2007]
∂x2 ∂y2 ∂x∂y
y−3x
, find x ∂u + y ∂u .
3
8. If u = x4 log 3 y+3x ∂x ∂y
3 y − 3 x
Hint : 3 y + 3 x is of degree 0 and u is of degree 4.
9. If z is a homogeneous function of degree n in x and y, and z is a function of u as z = f(u)
then
∂2 u ∂2 u ∂2 u x f (u)
x2 2 + 2xy + y2 2 = g(u) ( g'(u) − 1) , where g(u) =
∂x ∂x∂y ∂y f'(u)
288 Engineering Mathematics through Applications
4.5 TOTAL DERIVATIVES
Derivatives of Composite and Implicit Functions (Chain Rule)
Let –z = f(x, y) be a function of two independent variables x and y which are themselves
functions of one independent variable t, say x = φ (t), y = ψ(t) with an assumption that these
functions are differentiable. Assume that fx and fy are continuous functions of x and y.
Without actually substituting the values of x and y in f(x, y), we can find
dz ∂ z dx ∂ z dy d z ∂ f dx ∂ f dy
= + or = + …(1)
dt ∂ x dt ∂ y dt dt ∂ x dt ∂ y dt
Proof: If t is given an increment ∆t, then x, y and –z will have corresponding increments as ∆x,
∆y and ∆z, so that
∆z = f(x + ∆x, y + ∆y) – f (x, y)
= f(x + ∆x, y + ∆y) – f(x, y + ∆y) + f(x, y + ∆y) – f(x, y)
Dividing both sides by ∆t,
∆z f (x + ∆x, y + ∆y) − f (x, y + ∆y) ∆x f (x, y + ∆y) − f (x, y) ∆y
= ⋅ + ⋅
∆t ∆x ∆t ∆y ∆t
Taking limits as ∆t → 0, ∆x and ∆y also tends to zero.
∆z dz f (x + ∆x, y + ∆y) − f (x, y + ∆y) ∆x
∴ Lt = = Lt Lt ⋅ ∆t
∆t → 0 ∆t dt ∆x→ 0 ∆x
f (x, y + ∆y) − f (x, y) ∆y
+ Lt Lt ⋅ ∆t …(2)
∆y → 0 ∆y
The rules given in equations (1) and (2) are known as chain rules for functions defined
along the path.
These rules can be extended to a function of n variables
z = φ(x1, x2, … , xn) with x1 = φ1(t), x2 = φ2(t), …, xn = φn(t)
dz ∂f dx1 ∂f dx2 ∂f dxn
such that = + +…+ …(3)
dt ∂x1 dt ∂x2 dt ∂xn dt
Differentiation of Implicit Functions
If the function –z = f(x, y) = 0 defines implicitly a function y = φ(x) of one independent variable
x, then equation (1) reduces to
∂f
df ∂f ∂f dy dy ∂ ∂f
0= = + or =− x, provided ≠0 …(4)
dx ∂x ∂y dx dx ∂f ∂y
∂y
Further, if the function f(x, y, z) = 0 defines one of the variables x, y, z implicitly in terms
of the other two variables, then
∂f ∂f ∂f
df = dx + dy + dz = 0 …(5)
∂x ∂y ∂z
Partial Derivatives and their Applications 289
(i) If we take y = constant implying dy = 0, (5) gives
∂f
∂f ∂f ∂x
dx + dz = 0 or dz
=−
∂x ∂z dx y ∂f …(5a)
∂z
(ii) If we take x = constant implying dx = 0, (5) gives
∂f
∂f ∂f ∂y
dz
dy + dz = 0 or = −
∂y ∂z dy x ∂f …(5b)
∂z
(iii) If we take z = constant implying dz
– = 0, (5) gives
∂f
∂f ∂f dx ∂y
dx + dy = 0 or = −
∂x ∂y dy z ∂f …(5c)
∂x
Multiplying (5a), (5b), (5c), we get
dx dy dz ∂x ∂y ∂z
dy dz dx = −1 or ∂y ∂z ∂x = −1 …(5d)
z x y z x y
−1
du
Example 27: If u = sin–1(x – y), x = 3t and y = 4t3, show that = 3(1 − t2 ) 2 [PTU, 2005]
dt
Solution: Given, u = sin –1(x – y), so that du = ∂u dx + ∂u
dy
…(1)
dt ∂x dt ∂y dt
∂u 1 1
= =
Now, ∂x 1 − x − y + 2xy
2 2
1 − (3t) − (4t3 )2 + 2(3t)(4t3 )
2
1 1
= = , …(2)
(1 − 4t ) (1 − t ) (1 − 4t ) (1 − t2 )
2 2 2 2
∂y 1 1
= = ,
∂y 1 − x − y + 2xy (1 − 4t ) 1 − t2
2 2 2 …(3)
dx dy
= 3, = 12 t2. (from given) …(4)
dt dt
dx dy
On substituting values of ux, uy, , in (1), we get
dt dt
du 1 3(1 − 4 t2 ) −
1
= [3 − 12 t2
] ≡ = 3(1 − t2 ) 2 .
dt (1 − 4t ) 1 − t
2 2 1
(1 − 4 t2 )(1 − t2 )2
Example 28: Find the total differential coefficient of x2y with respect to x when x and y
are connected by the relation x2 + xy + y2 = 1. [KUK, 2008; NIT Kurukshetra, 2008]
dz
Solution: Let z = x2y, then in this problem we need to find .
dx
290 Engineering Mathematics through Applications
dz ∂z dx ∂z dy dy
= + = 2xy + x2 …(1)
dx ∂x dx ∂y dx dx
Further, variables x and y are connected by an implicit relation, φ = x + xy + y = 0
2 2
comparable to φ(x, y) = 0
∂φ
2x + y
= − ∂x = −
dy …(2)
so that
dx ∂φ 2y + x
∂y
dy
On substituting value of from (2) in (1), we get
dx
dz 2x + y
= 2 xy − x2 .
dx 2y + x
∂f ∂φ dz ∂f ∂φ
Example 29: If f(x, y) = 0, φ(y, z) = 0, show that ∂y ∂z dx = ∂x ∂y .
Solution: As f (x, y) = 0 is an implicit relation between x and y which defines a differential
coefficient of y with respect to x,
∂f
dy ∂x
=−
dx ∂f …(1)
∂y
∂φ
∂y
dz
Likewise for φ(y, z) = 0, dy = − …(2)
∂φ
∂z
∂φ ∂f
dz dz dx ∂y
dz ∂y ∂f ∂φ dz ∂f ∂φ
write, = ⋅ ⇒ − =− ⇒ ⋅ = .
dy dx dy ∂φ dx ∂f ∂y ∂z du ∂x ∂y
∂z ∂x
Example 30: At a given instant that sides of a rectangle are 4 ft and 3 ft respectively and
they are increasing at the rate of 1.5 ft/s and 0.5 ft/s respectively, find the rate at which the
area is increasing at that instant.
dy
Solution: Given x = 4 ft, y = 3 ft, and dx = 1.5 ft/s, = 0.5 ft/s
dt dt
Here z = f(x, y) = x · y (area of the rectangle)
dz ∂z dx ∂z dy ∂z ∂ ∂z
so that = + = y (1.5) + x (0.5) , as = (xy) = y and = x
dt ∂x dt ∂y dt ∂x ∂x ∂y
= 3(1.5) + 4(0.5) = 4.5 + 2.0 = 6.5 ft/s.
Example 31: If z = (2xy2 – 3x2y) and if x increases that the rate of 2 cm/s and it passes
through x = 3 cm, show that if y is passing through the value y = 1 cm, y must be decreasing
2
at the rate 2 15 cm/s in order that z shall remain constant.
Partial Derivatives and their Applications 291
dz ∂z dx ∂z dy
Solution: For z = (2xy2 – 3x2y), = +
dt ∂x dt ∂y dt
dx dy
= (2y2 − 6xy)+ (4xy − 3x2 )
dt dt
dx dz dy
Given = 2, so that = (2y2 − 6xy)2 + (4xy − 3x2 )
dt dt dt
dz dy
when x = 3, y = 1, then = −32 − 15
dt dt
dy
In order that –z remains constant, dz = 0 implies 0 = −32 − 15
dt dt
dy 32 2
or = =2 cm/s
dt 15 15
d2 y q2 r − 2 pqs + p2t
Example 32: If f(x, y) = 0, show that = − , [KUK, 2006]
dx2 q3
∂f ∂f ∂2 f ∂2 f ∂2 f
where p = , q= , r = 2 , s= , t= 2
∂x ∂y ∂x ∂x ∂y ∂y
Solution: For any function f(x, y), we have
d ∂ dx ∂ dy
f (x, y) = f (x, y) + f (x, y)
dt ∂x dt ∂y dt
In particular, when t = x,
d dy
f (x, y) = fx (x, y) ⋅ 1 + fy (x, y) …(1)
dx dx
Identically,
d dy
f (x, y) = fxx(x, y) + fxy(x, y) …(2)
dx x dx
d dy
and f (x , y) = fyx (x , y) + fyy (x , y) …(3)
dy y dx
As we are given that f(x, y) = 0, i.e. f(x, y) is an implicit function, therefore from (1), we get
d dy dy f (x, y)
0= f (x, y) = fx (x, y) + fy (x, y) or =− x …(4)
dx dx dx fy (x, y)
Thus,
d ( , ) d ( , )
d2 y d dy d fx fy fx x y − fx fy x y
dx dx
2 = dx =
dx dx fy
− =− 2
dx fy
Applying U formula for derivatives
V
=−
{
fy fxx + fyx
dy
dx} {
− fx fyx + fyy
2
dy
dx}, [on using (2),(3)]
fy
292 Engineering Mathematics through Applications
f f
fy fxx + fxy − x − fx fxy + fyy − x
y
f
fy
=− , [on using(4)]
fy2
=−
{ } {
fy fy fxx − fx fxy − fx fy fxy − fx fyy },
fy3
d2 y fy2 fxx − 2 fx fy fxy + fx2 fyy
= − …(5)
dx2 fy3
d2 y q2r − 2pqs + p2t
2 =−
or , (on using given notations) …(6)
dx q3
Also, on cross multiplication in (5), we get
d2 y
fy3 = 2 fy fx fxy − fx2 fyy − fy2 fxx …(7)
dx2
Hence the proof.
d2 y
Example 33: xn + yn = an, find .
dx2
Solution: xn + yn = an is an implicit relation for which
dy f
=− x =−
p d2 y q2 r − 2pqs + p2 t
dx fy q and 2 = − .
dx q3
Here p = fx = nxn –1, q = fy = nyn –1,
r = fxx = n(n – 1)xn –2, s = fxy = 0, t = n(n – 1)yn – 2.
d2 y 1 2
2 = − 3 (q r − 2pqs + p t)
2
∴ dx q
1
= n2 y2n − 2 ⋅ n(n − 1)xn − 2 − 2n2 xn −1 yn − 1 ⋅ 0 + n2 x2n − 2n(n − 1)yn − 2
n3 y3n − 3
n3 (n − 1) n − 2 2n − 2
= x ⋅ y + yn − 2 x2n − 2
n3 y3n− 3
−(n − 1)xn −2 yn −2 n
= y + xn
y3n − 3
xn − 2
= −(n − 1) an ⋅ .
y2n −1
dx dy dz
Example 34: If x2 + y2 + z2 – 2xyz = 1, show that + + =0
1−x 2
1−y 2
1 − z2
[NIT Kurukshetra, 2009]
Partial Derivatives and their Applications 293
Solution: As the given relation x2 + y2 + z2 – 2xyz = 1 is an implicit function comparable to
f(x, y, z) = c, has its total differential coefficient equal to zero.
i.e. df = fx dx + fy dy + fzdz = 0
(x – yz)dx + (y – zx)dy + (z – xy)dz = 0 (on simplification) …(1)
2 2 2 2 2 2
Rewriting the given equation, x + y + z – 2xyz = 1 as x – (2yz)x – (1 – y – z ) = 0
which is comparable to a general quadratic equation, ax2 + bx + c = 0 with a = 1, b = –2yz,
c = –(1 – y2 – z2).
−b ± b2 − 4ac 2yz ± 4y z + 4(1 − y − z )
2 2 2 2
∴ x= =
2a 2
2yz ± 2 (1 − y2 ) − z2(1 − y2 )
x= = yz ± (1 − y2 )(1 − z2 )
2
or (x − yz) = (1 − y2 )(1 − z2 ) , (Taking the positive sign only) …(2)
Similarly, (y − zx) = (1 − z2 )(1 − x2 ) , …(3)
and (z − xy) = (1 − x2 )(1 − y2 ) …(4)
On using above results in equation (1),
(1 − y2 )(1 − z2 ) dx + (1 − x2 )(1 − z2 ) dy + (1 − x2 )(1 − y2 ) dz = 0
dx dy dz
or + + =0
1 − x2 1 − y2 1 − z2
(Dividing throughout by (1 − x2 )(1 − y2 )(1 − z2 ) )
Example 35: If the three thermodynamical variable P, V, T are connected by the relation
f(P, V, T) = 0, show that
∂P ∂T ∂V
⋅ ⋅ = − 1.
∂T V ∂V P ∂P T
Solution: The function f(P,V,T) = 0 defines an implicit relation on the three variables
P(pressure), V(volume), T(temperature) such that each of them is a function of the other two.
∴ df = fx dP + fy dV + fz dT = 0
∂f
∂P = − ∂T
(i) if V is kept constant, i.e. dV = 0, then
∂T V ∂f …(1)
∂P
(i.e. change in one variable with respect to the other keeping the 3rd constant)
∂f
(ii) if P is kept constant, i.e. dP = 0, then
∂ T = − ∂V
∂V P ∂f …(2)
∂T
294 Engineering Mathematics through Applications
∂f
∂P
∂V
(iii) if T is kept constant, i.e. dT = 0, then =−
∂P T ∂f …(3)
∂V
∂P ∂T ∂V
On multiplying the three, ∂T ∂V ∂P = −1
V P T
Note: In case of implicit function with more than 2 variables, the differential coefficients involved are obviously
partial in nature instead of total.
Example 36: The pressure p, volume v and absolute temperature T of a gas are connected
by the relation pv = RT, where R is constant. If Q is a function of the state of a gas such
∂Q ∂Q ∂Q
that = p then show that = + R. Also find ∂Q .
∂v v,T ∂ T p,T ∂ T v,T ∂p p,T
Solution: We had the relation pv = RT, …(1)
connecting pressure p, volume v and temperature T.
We can always consider Q the state of a gas as function of above three independent variables
p, v, T.
∂Q
Here in the given problem, clearly the notation, ∂v defines that Q is a function of
v,T
two independent variables v and T, and its partial derivative is taken with v keeping T
constant.
For Q(v, T),
∂Q ∂Q
dQ = dv +
∂v ∂T
dT …(2)
Likewise for Q(p, T),
∂Q ∂Q
dQ = dp +
∂T p,T
dT
∂ p p, T …(3)
On equating the two values, we get
∂Q ∂Q ∂Q ∂Q
∂v dv + ∂T dT = ∂p dp + ∂T dT …(4)
v, T v,T p, T p,T
Now the gas relation pv = RT gives
p dv + v dp = R dT …(5)
∂Q ∂Q
or ∂v dv + v dp = R dT given ∂v = p
v, T v, T
∂Q
∂v dv = (R dT − v dp) …(6)
v, T
On substituting (6) into (4), we get
∂Q ∂Q ∂Q
(R dT − v dp) + dT = dp +
∂T v,T ∂p p,T ∂T p,T
dT
Partial Derivatives and their Applications 295
∂Q ∂Q ∂Q
or R + ∂T dT − v dp = ∂p dp + ∂T dT …(7)
v, T p, T p, T
On both sides, comparing co-efficients of dT and dp, we get
∂Q = + ∂Q and ∂Q = −
∂T R
∂T v,T ∂p p,T
v
p, T
Hence the proof.
ASSIGNMENT 4
du
1. If u = x2 + y2 + z2 with x = e2t , find and verify the result by direct substitution.
dt
y = e2t cos 3t,
z = e2t sin 3t
dy
2. Given a relation between x and y, (cos x)y = (sin y)x, find . [KUK, 2005]
dx
du
3. If u = sin(x2 + y2), where a2x2 + b2y2 = c2, find .
dx
du
4. If u = x logxy where x3 + y3 + 3xy = 1, find .
dx
5. If x increases at the rate of 2cm/sec at the instant when x = 3cm and y = 1cm, at what rate
must y be changing in order that the function (2xy – 3x2y) shall neither increasing nor
du
decreasing, i.e. = 0.
dt
∂f ∂φ ∂f ∂φ
6. If the curves f(x, y) = 0 and φ(x, y) = 0 touch, show that at the point of contact ∂y ∂x = ∂x ∂y
[Hint: The two curves touch means, at the point of contact, the tangent is common.]
∂y ∂z
7. Given that z3 + xy – y2z = 6, obtain expressions for , in terms of x, y, z and find
∂x ∂x
∂y fx ∂z fx
their values at the point (0, 1, 2). Hint : If (x, y, z) = 0, then ∂x = − f , ∂x = − f
z y y z
8. If ax2 + by2 + cz2 = 1 and lx + my + nz = 0, prove that
dx dy dz
= =
bny − cmz clz − anx amx − bly
4.6 CHANGE OF VARIABLES: FUNCTIONS DEFINED ON SURFACES
Suppose that u = f(x, y) is a function of two independent variables (x, y) and x, y are further
dependent on two independent variables (s, t) given by x = φ(s, t) and y = ψ(s, t), then by
chain rule,
∂ u ∂ u ∂ x ∂ u ∂y
= + …(1)
∂s ∂x ∂s ∂y ∂s
296 Engineering Mathematics through Applications
∂ u ∂ u ∂ x ∂ u ∂y
and = + …(2)
∂t ∂x ∂t ∂y ∂t
Here ordinary derivatives have been replaced by partial derivatives, since x and y are
further functions of s and t (instead of one variable t as in the case of total derivatives, Article 4.5,
equation (1)).
∂u ∂u
On solving (1) and (2) as simultaneous equations in and ∂y , we get their values in
∂x
∂u ∂ u
terms of , , s, t which precisely means that if s and t are given as functions s = ξ(x, y)
∂s ∂t
and t = η(x, y), then we get
Dependent
∂u ∂u ∂s ∂u ∂t u = f(x, y)
= + …(3)
variable
∂x ∂s ∂x ∂t ∂x
∂u
∂u
∂u ∂u ∂s ∂u ∂t ∂x
and = + …(4)
∂y
∂y ∂s ∂y ∂t ∂y
as shown in chain rule Fig. 4.1. X Intermediate
Y
The chain rule may be extended to a function, variable
u = f(x, y, z) of three independent variable which
are further functions of the independent variables ∂x ∂y
∂s ∂s
s and t as follows:
∂u ∂u ∂x ∂u ∂y ∂u ∂z
= + + …(5)
∂s ∂x ∂s ∂y ∂s ∂z ∂s
Fig. 4.1
∂u ∂u ∂x ∂u ∂y ∂u ∂z
and = + + …(6)
∂t ∂x ∂t ∂y ∂t ∂z ∂t
as shown by chain rule, Fig. 4.2.
Dependent u = f (x , y , z )
u = f (x , y , z )
variable
∂u ∂u ∂u ∂u ∂u
∂x ∂y ∂z ∂u
∂x ∂z
∂y
y y
X Z Intermediate X Z
variable
∂y
∂y ∂x ∂t
∂x ∂t ∂z
∂s ∂z
∂s ∂t
∂s
s Independent L
variable
Fig. 4.2 (i ) Fig. 4.2 ( ii )
Partial Derivatives and their Applications 297
∂z ∂z ∂z ∂z
Example 37: If z = f(x, y) and x = eu + e− v prove that ∂u − ∂v = x ∂x − y ∂y . [VTU, 2006]
,
y = eu − ev
x(u, v) = eu + e−v
Solution: From given we get
y(u, v) = e−u − ev
∂x ∂x ∂y ∂y
= eu , = − e− v and = − e−u , = −ev
∂u ∂v ∂u ∂v
∂z ∂z ∂x ∂z ∂y ∂z ∂z
∴ = + = eu − e −u …(1)
∂u ∂x ∂u ∂y ∂u ∂x ∂y
∂z ∂z ∂x ∂z ∂y ∂z v ∂z
and = + = − e−v −e …(2)
∂v ∂x ∂v ∂y ∂v ∂x ∂y
∂z ∂z − v ∂z v ∂z ∂z ∂z
(1) – (2) gives, ∂u − ∂v = ( e + e ) ∂x − (e − e ) ∂y = x ∂x − y ∂y .
u −u
Example 38: For
x = r cosθ
y = r sinθ }
, z = f (x, y) has continuous partial derivatives, show that
∂z 1 ∂z
(i) = fx cosθ + fy sinθ and = − fx sinθ + fy cosθ
∂r r ∂θ
2 2
∂z 1 ∂z
(ii) ∂r + r 2 ∂θ = fx + fy .
2 2
Solution: For z = f(x, y),
∂ z ∂f ∂ x ∂f ∂ y
= + = fx cos θ + fy sin θ …(1)
∂r ∂x ∂r ∂y ∂r
∂x ∂y
since x = r cos θ gives ∂r = cos θ and y = r sin θ gives ∂r = sin θ
θ θ
∂z ∂f ∂x ∂f ∂y
and = + = f (−r sin θ) + fy (r cos θ)
∂θ ∂x ∂θ ∂y ∂θ x
∂x ∂y
since x = r cos θ gives ∂θ = −r sin θ and y = r sin θ gives ∂θ = r cos θ
r r
1 ∂z
or = − fx sin θ + fy cos θ …(2)
r ∂θ
Squaring and adding (1) and (2),
( ) ( )
2 2
∂z + 1 ∂z = f cos θ + f sin θ 2 + − f sin θ + f cos θ 2
∂r r2 ∂θ x y x y
298 Engineering Mathematics through Applications
= (fx2 cos2θ + fy2 sin2θ + 2fx fysinθ cosθ) + (fx2 sin2θ + fy2 cos2θ – 2fxfy sinθ cosθ)
= (cos2θ + sin2θ) fx2 + (cos2θ + sin2θ) fy2 = fx2 + fy2.
1 ∂u 1 ∂u 1 ∂u
Example 39: If u = f (2x – 3y, 3y – 4z
–, 4z
– – 2x), prove that + + =0
2 ∂x 3 ∂y 4 ∂z
[NIT Kurukshetra, 2002, 2004; Raipur, 2005; UP Tech, 2006]
Solution: Consider u = f (2x – 3y, 3y – 4z, 4z – 2x) = f (r, s, t)
where r = 2x – 3y, s = 3y – 4z, t = 4z – 2x
∂u ∂u ∂r ∂u ∂s ∂u ∂t
So that = + +
∂x ∂r ∂x ∂s ∂x ∂t ∂x
∂u ∂u ∂u ∂u ∂u
= 2+ 0+ × −2 = 2 −2 …(1)
∂r ∂s ∂t ∂r ∂t
∂u ∂u ∂r ∂u ∂s ∂u ∂t
Similarly, = + +
∂y ∂r ∂y ∂s ∂y ∂t ∂y
∂u ∂u ∂u ∂u ∂u
= × −3 + 3+ 0 = −3 +3 …(2)
∂r ∂s ∂t ∂r ∂s
∂u ∂u ∂r ∂u ∂s ∂u ∂t
and = + +
∂z ∂r ∂z ∂s ∂z ∂t ∂z
∂u ∂u ∂u ∂u ∂u
= 0+ × −4 + 4 = −4 +4 …(3)
∂r ∂s ∂t ∂s ∂t
1 ∂u 1 ∂u 1 ∂u
On using (1), (2) and (3), we see that 2 ∂x + 3 ∂y + 4 ∂z = 0
Example 40: If u = f (x2 + 2yz, y2 + 2zx), prove that
∂u ∂u ∂u
(y2 − zx) + (x2 − yz) + (z2 − xy) =0
∂x ∂y ∂z
Solution: Consider u = f (x2 + 2yz, y2 + 2zx) = f (r, s), r = x2 + 2yz and s = y2 + 2zx …(1)
Differentiating u(x, y, z) partially with respect to x, y, z respectively as below:
∂ u ∂f ∂ r ∂f ∂ s
= +
∂x ∂r ∂x ∂s ∂x
∂ u ∂f ∂ r ∂f ∂ s
= + …(2)
∂y ∂r ∂y ∂s ∂y
∂ u ∂f ∂ r ∂f ∂ s
= +
∂z ∂r ∂z ∂s ∂z
∂r ∂r ∂r
and = 2x, = 2z, = 2y;
∂x ∂y ∂z
∂s ∂s ∂s …(3)
= 2z, = 2y, = 2x;
∂x ∂y ∂z
Partial Derivatives and their Applications 299
∂r ∂r ∂r ∂s ∂s ∂s
∴ On substituting values of ∂x , ∂y , ∂z , ∂x , ∂y , ∂z in equation (2), we get
∂u
= ( fr ⋅ 2x + fs ⋅ 2z)
∂x
∂u
= ( fr ⋅ 2z + fs ⋅ 2y)
∂y
…(4)
∂u
= ( fr ⋅ 2y + fs ⋅ 2x)
∂z
Using determinant method, eliminate fr and fs from (4),
∂u
2x 2z
∂x
∂u
2z 2y =0
∂y
∂u
2y 2x
∂z
∂u ∂u ∂u ∂u ∂u
or 2x 2y
∂z
− 2x + 2z 2y
∂y ∂y
− 2z +
∂z ∂x
( 2z ⋅ 2x − 2y ⋅ 2y ) = 0
On re-arrangement of terms, we get the desired result.
∂u ∂u ∂u
(y2 − zx) + (x2 − yz) + (z2 − xy) = 0.
∂x ∂y ∂z
∂2 u ∂2 u
Example 41: Transform the equation + = 0 into polar co-ordinates.
∂x2 ∂y2
[UP Tech. 2004; KUK, 2007]
∂2 u ∂2 u ∂u ∂u
Solution: For finding 2 and ∂y2 , we need to find first ∂x and ∂y .
∂x
We know that x = r cos θ
so that
y = r sin θ } 1
y
implies r = (x2 + y2 )2 , θ = tan−1
x
…(1)
∂r 1 1 x
= ⋅ 1 ⋅ 2x = = cos θ,
∂x 2 r
(x2 + y2 )2
∂r 1 1 y
= ⋅ 1 ⋅ 2y = = sin θ,
∂y 2 r
(x2 + y2 )2
∂θ 1 ∂ y x2 y y sin θ …(2)
= = − = − = − ,
∂x y ∂ x x x2 + y2 x2 r
2
r2
1 +
x
∂θ 1 ∂ y x2 1 = x = cos θ ,
= = ⋅
∂y y
2
∂ y x x2
+ y2
x r2
r
1+
x
300 Engineering Mathematics through Applications
Now,
∂u ∂u ∂r ∂u ∂θ
= +
∂x ∂r ∂x ∂θ ∂x
and ∂u ∂u ∂r ∂u ∂θ …(3)
= +
∂y ∂r ∂y ∂θ ∂y
∂u ∂u
On using equations (2), above terms and ∂y become as
∂x
∂u ∂u sin θ ∂u ∂ sin θ ∂
= cos θ − = cos θ − u …(4)
∂x ∂r r ∂θ ∂r r ∂θ
∂u ∂u cos θ ∂u ∂ cos θ ∂
and = sin θ + = sin θ + u …(5)
∂y ∂r r ∂θ ∂r r ∂θ
∂ ∂
From (4) and (5), we get the values of the operators and ∂y as:
∂x
∂ ∂ sin θ ∂ ∂ ∂ cos θ ∂
= cos θ − and = sin θ − …(6)
∂x ∂r r ∂θ ∂y ∂r r ∂θ
∂2u ∂ ∂u ∂2 u ∂ ∂u
Now to find , apply on and similarly, to find 2 , apply ∂y on ∂y ,
∂x2 ∂x ∂x ∂ y
∂2u ∂ sin θ ∂ ∂u sin θ ∂u
∴ = cos θ − cos θ −
∂x 2 ∂r r ∂θ ∂r r ∂θ
∂ ∂u ∂ sin θ ∂u sin θ ∂ ∂u sin θ ∂ sin θ ∂u
= cos θ cos θ + cos θ − − cos θ − −
∂r ∂r ∂r r ∂θ r ∂θ ∂θ r ∂θ r ∂θ
∂2u 2 ∂ u
2
sin θ cos θ ∂2u sin θ cos θ ∂u sin2 θ ∂u sin2 θ ∂2u
⇒ 2 = cos θ 2 −2 ∂r∂θ
+2
∂θ
+
r ∂r
+ 2 …(7)
∂x ∂r r r2 r ∂θ2
∂2u ∂ cos θ ∂ ∂u cos θ ∂u
and ∂y2 = sin θ ∂r + r ∂θ sin θ ∂r + r ∂θ
∂ ∂u ∂ cos θ ∂u cos θ ∂ ∂u cos θ ∂ cos θ ∂u
= sin θ sin θ + sin θ + sin θ +
∂r ∂r ∂r r ∂θ r ∂θ ∂r r ∂θ r ∂θ
∂2u ∂2u sin θ cos θ ∂2u 2 sin θ cos θ ∂u cos2 θ ∂u cos2 θ ∂2u
⇒ = sin2 θ 2 + 2 − + + …(8)
∂y 2
∂r r ∂r∂θ r2 ∂θ r ∂r r2 ∂θ2
On adding (7) and (8), we get
∂2 u ∂2 u ∂2 u 1 ∂u 1 ∂2 u
+ = + +
∂x2 ∂y2 ∂r2 r ∂r r2 ∂θ2
Hence the desired result.
Partial Derivatives and their Applications 301
Example 42: If u = f(x, y) and x = r cosθ prove that
y = r sinθ
, }
2 2 2 2
∂u ∂u ∂u 1 ∂u
(i) ∂x + ∂y = ∂r + 2 ∂θ ,
r
∂2 u ∂2 u ∂2 u 1 ∂2 u 1 ∂u
(ii) ∂x2 + ∂y2 = ∂r 2 + r2 ∂θ2 + r ∂r [Pb. Univ, 2002]
Solution: This example is an another way of understanding the previous example. In the
∂2u ∂2u ∂2 u 1 ∂2 u 1 ∂u
previous example we have just proved that 2 + 2 = 0 transforms to 2 + 2 2 + r ∂r = 0
∂x ∂y ∂r r ∂θ
∂2u ∂2u
Means the expression, + which is in cartisian coordinate system has value
∂ x2 ∂ y 2
∂2u 1 ∂2u 1 ∂u
+ + in (r, θ) system.
∂r2 r2 ∂θ2 r ∂θ
In the previous example, under equation (3), we had
∂u
∂x
= cos θ(∂u sin θ ∂u
∂r
−
r ∂θ and ) ∂u
= sin θ
∂y
∂u cos θ ∂u
∂r
+
r ∂θ
On squaring and adding these two results, we get
2 2 2 2
∂u + ∂u = cos θ ∂u − sin θ ∂u + sin θ ∂u + cos θ ∂u
∂x ∂y ∂r r ∂θ ∂r r ∂θ
∂u
2
sin2 θ ∂u
2
sin θ cos θ ∂u ∂u
= cos2 θ + 2 − 2
∂r r ∂θ r ∂r ∂θ
∂u
2
cos2 θ ∂u
2
sin θ cos θ ∂u ∂u
+ sin2 θ + 2 + 2
∂ r r ∂θ r ∂r ∂θ
2 2 2 2
∂u ∂u ∂u 1 ∂u
= (cos2 θ + sin2 θ) + 2 (cos2 θ + sin2 θ) = + 2
1
∂r r ∂θ ∂r r ∂θ
φ, y = r cosθ
θ cosφ
Example 43: Given that u(x, y, z) where x = r cosθ θ sinφ
φ and z = r sinθ
θ, find
∂u ∂u
and . [NIT Kurukshetra, 2008]
∂θ ∂φ
Solution: Here by change of variable, we have
∂u ∂u ∂x ∂u ∂y ∂u ∂z
= + + …(1)
∂θ ∂ x ∂θ ∂ y ∂θ ∂z ∂θ
∂x
= −r sin θ cos φ,
x = r cos θ cos φ, ∂θ
∂y
Now y = r cos θ sin φ, ⇒ = −r sin θ sin φ,
z = r sin θ ∂θ …(2)
∂z
= r cos θ
∂θ
302 Engineering Mathematics through Applications
Using (2), we get
∂u
= f'(x2 + y2 + z2 ) ⋅ 2x( −r sin θ cos φ) + f'(x2 + y2 + z2 ) ⋅ 2y(−r sin θ sin φ)
∂θ
+ f'(x2 + y2 + z2 ) ⋅ 2 z(r cos θ) …(3)
Simplifying the above expressions in one system of variables either express (r, θ, φ) or
(x, y, z), say.
∂u
∴ = f'(r2 ) ⋅ 2r cos θ cos φ (−r sin θ cos φ) + f'(r2 )2r cos θ sin φ(−r sin θ⋅ sin φ)
∂θ
+ f'(r2 )2r sin θ(r cos θ)
= 2r2f’(r2) [–cosθ sinθ cos2φ – cosθ sinθ sin2φ + sinθ cosθ]
= 2r2f’(r2) [–cosθ sinθ (cos2φ + sin2φ) + cosθ sinθ]
= 2r2f’(r2) [–cosθ sinθ + cosθ sinθ] = 0
∂u ∂u ∂x ∂u ∂y ∂u ∂z
Similarly = + + …(4)
∂φ ∂ x ∂φ ∂ y ∂φ ∂z ∂φ
∂x
= −r cos θ sin φ
∂φ
∂y
whever = r cos θ cos φ
∂φ …(5)
∂φ
=0
∂z
∂u
∴ = f '(x2 + y2 + z2 ) ⋅ 2x(−r cos θ sin φ) + f '(x2 + y2 + z2 ) · 2y(rcosθ cosφ
∂φ
+ f’(x2 + y2 + z2)2z = 0
= f ’(r2) 2 rcosθ cosφ (– rcosθ sinφ) + f ’(r2)2 rcosθ sinφ (rcosθ cosφ)
= 2r2 f ’(r2) [cosφ sinφ cos2θ – cosφ sinφ cos2θ] = 0
∂2 u ∂2 u ∂2 u
Example 44: If x + y = 2eθ cosφ
φ, x – y = 2ieθ sinφ
φ, show that + = 4 xy ∂ x2
∂θ2 ∂φ2
[UP Tech, 2002]
∂2u ∂2u ∂u ∂u
Solution: In order to find 2 and ∂φ2 , we need to find first ∂θ and ∂φ respectively.
∂θ
∂ u ∂ u ∂ x ∂ u ∂y
We know that = + …(1)
∂θ ∂x ∂θ ∂y ∂θ
∂ u ∂ u ∂ x ∂ u ∂y
and = + …(2)
∂φ ∂x ∂φ ∂y ∂φ
x + y = 2 eθ cos φ x = eθ+ i φ
Given implies …(3)
x − y = 2 i eθ sin φ y = eθ− i φ
Partial Derivatives and their Applications 303
∂x ∂x
So that = eθ + iφ ⋅ 1 = x; = eθ + iφ ⋅ i = ix;
∂θ ∂φ
∂y ∂y …(4)
= eθ − iφ ⋅ 1 = y; = eθ − iφ ⋅ − i = −iy
∂θ ∂φ
On using (4), (1) and (2) becomes
∂u ∂u ∂u ∂ ∂
= x + y = x +y u …(5)
∂θ ∂ x ∂y ∂x ∂y
∂u ∂u ∂u ∂ ∂
and = ix − iy = ix − iy u …(6)
∂φ ∂ x ∂y ∂ x ∂y
∂ ∂
Clearly from above two equations, the value of the operators and
∂θ ∂φ are
∂ ∂ ∂
= x +y
∂θ ∂ x ∂y
and ∂ ∂ ∂ …(7)
= ix − iy
∂φ ∂ x ∂ y
Therefore,
∂2u ∂ ∂u ∂ ∂ ∂u ∂u
2 = ∂θ ∂θ = x ∂x + y ∂y x ∂x + y ∂y
∂θ
∂ ∂u ∂ ∂u ∂ ∂u ∂ ∂u
= x x + x y + y x + y y
∂x ∂x ∂x ∂y ∂y ∂x ∂y ∂y
14444 4244444 3 14444 4244444 3
∂u ∂2 u ∂2u ∂2u ∂u ∂2 u
=x + x2 2 + xy + yx +y + y2 2
∂x ∂24444
x ∂x∂y ∂y∂x ∂y ∂3
y
14444 3 1444424444
∂2 u ∂2 u ∂2 u ∂u ∂u
= x2 2 + 2xy + y2 2 + x +y …(8)
∂x ∂x∂y ∂y ∂x ∂y
Likewise,
∂2u ∂ ∂u ∂ ∂ ∂u ∂u
= = ix − iy ix − iy
∂φ2 ∂φ ∂φ ∂x ∂y ∂x ∂y
∂2 u ∂2 u ∂2 u ∂u ∂u
= −x2 2 + 2xy − y2 2 − x +y …(9)
∂x ∂ x∂ y ∂y ∂ x ∂ y
On adding (8) and (9),
∂2u ∂2u ∂2u
+ 2 = 4xy
∂θ2
∂φ ∂ x∂ y
304 Engineering Mathematics through Applications
Example 45: If by the substitution u = x2 – y 2, v = 2xy, f(x,y) = θ (u,v). Show that
∂2 f ∂2 f ∂2 θ ∂2 θ
2 +
= 4(x2 + y2 ) 2 + 2 [Anna., 2003]
∂x ∂y2
∂x ∂y
∂2 f ∂2 f ∂f ∂f
Solution: In order to find and ∂y 2 , we need to find first and ∂y respectively.
∂x 2 ∂x
Thus,
∂f ∂f ∂u ∂f ∂v
= + …(1)
∂x ∂u ∂x ∂v ∂x
∂f ∂f ∂u ∂f ∂v
and = + …(2)
∂y ∂u ∂y ∂v ∂y
But from the given, u(x, y) = x2 – y2 and v(x, y) = 2xy,
∂u ∂u
= 2x, = −2y
∂x ∂y
we have ∂v ∂v …(3)
= 2y, = 2x
∂x ∂y
On using (3), equations (1) and (2) become
∂f ∂f ∂f
= 2x + 2y
∂x ∂u ∂v
∂f ∂f ∂f
= −2 y + 2x
and ∂y ∂u ∂v
∂f ∂ ∂ ∂ ∂ ∂θ ∂θ
Rewrite = 2x + 2y f (x, y) = 2x + 2y θ(u, v) = 2x + 2y …(4)
∂x ∂u ∂v ∂u ∂v ∂u ∂v
∂f ∂ ∂ ∂ ∂ ∂θ ∂θ
and = −2y + 2x f (x, y) = −2y + 2x θ(u, v) = −2y + 2x
∂y ∂u ∂v ∂u ∂v ∂u ∂v
…(5)
From the above equations (4) and (5), we see that
∂f ∂θ ∂θ ∂f ∂θ ∂θ
= 2x + 2y = −2y + 2x
∂x ∂u ∂v ∂y ∂u ∂v
∂ ∂
and respectively
∂ ∂ ∂ ∂
= 2x + 2y = −2y + 2x
∂x ∂u ∂v ∂y ∂u ∂v
Now
∂2 f
∂x
∂ ∂f
( ∂ ∂
2 = ∂x ∂x = 2 x ∂u + 2 y ∂v )(2x ∂∂θu + 2y ∂θ∂v )
∂2 f 2 ∂ θ
2
∂2θ ∂2θ 2 ∂ θ
2
or 2 = 4x 2 + 4yx ∂v∂u + 4xy ∂u∂v + 4y …(6)
∂x ∂u ∂v2
Partial Derivatives and their Applications 305
∂2 f ∂ ∂f ∂ ∂ ∂θ ∂θ
Similarly = = −2y + 2x −2y + 2x
∂y2 ∂y ∂y ∂u ∂v ∂u ∂v
∂2 f 2 ∂ θ
2
∂2θ ∂2θ 2 ∂ θ
2
or 2 = 4y 2 − 4yx ∂v∂u − 4xy ∂u∂v + 4x …(7)
∂y ∂u ∂v2
Add (6) and (7), we get the desired result
∂2 f ∂2 f 2 ∂ θ
2
∂2 θ
2 + 2 = 4(x + y ) + 2
2
∂x ∂y ∂u 2
∂v
ASSIGNMENT 5
1. If u =f(r, s), r = x + at, s = y + bt, and x, y, t are independent variables, show that
∂u ∂u ∂u
=a +b .
∂t ∂x ∂y
∂u ∂u ∂u
2. If u = F(x – y, y – z, z – x), prove that ∂x + ∂y + ∂z = 0
x y z ∂u ∂u ∂u
3. If u = f(r, s, t) and r = , s = , t = , prove that x +y +z = 0 [Karnataka, 2006]
y z x ∂x ∂y ∂z
4. If x = u + v + w, y = uw + vw + uv and F is a function of x, y, z, show that
∂F ∂F ∂F ∂F ∂F ∂F
u +v +w =x + 2y + 3z .
∂u ∂v ∂w ∂x ∂y ∂z
∂2
5. If x = r cosθ, y = r sinθ, z = f(x, y), prove that ∂x∂y (r cos n θ) = −n(n − 1)r sin(n − 2) θ
n n−2
∂ ∂ sin θ ∂ ∂ ∂ cos θ ∂
Hint : Follow Ex. ∂x = cos θ ∂r − r ∂θ , ∂y = sin θ ∂r + r ∂θ
∂2u ∂2u −2r ∂ u
2
∂2u
6. If x = ercosθ, y = ersinθ; prove that 2 + 2 = e
∂r
+ 2
∂x ∂y 2
∂θ
∂ ∂ ∂ ∂ ∂ ∂
Hint: Use ∂r = x ∂x + y ∂y , ∂θ = −y ∂x + x ∂y
7. If z is a function of x and y and u = lx + my and v = ly – mx, then prove that
∂2 z ∂2 z 2 ∂ z
2
∂2 z
2 + 2 = (l + m ) + 2
2
∂x ∂y ∂u 2
∂v
∂y ∂x ∂y
8. If w = u(x, y), where x = x(u, v), y = y(u, v), ∂x = and = − , show that
∂u ∂v ∂v ∂u
∂2 w ∂2 w ∂2 w ∂2 w ∂x ∂x
2 2
2 + 2 = 2 + 2 ∂u + ∂v .
∂u ∂v ∂x ∂y
9. If u = f(x, y, z) and x = r sinθ cosφ, y = r sinθ sinφ, z = r cosθ, then show that
2 2 2 2 2 2
∂f ∂f ∂f ∂f 1 ∂f 1 ∂f
+ + = + 2 + 2
∂x ∂y ∂z ∂r r ∂θ r sin 2 θ ∂φ
[Hint: Follow Example 42 (ii), Section 4.6]
306 Engineering Mathematics through Applications
10. By changing the independent variables u and v to x and y by means of the relation
∂2 z ∂2 z
x = (u cos α – v sin α), y = (u sin α + v cos α), show that + transforms to
∂ u2 ∂ v2
∂2 z ∂2 z
+ . [NIT Kurukshetra, 2007]
∂ x2 ∂ y2
4.7 JACOBIANS: CO-EFFICIENTS OF TRANSFORMATION OF SYSTEM OF
VARIABLES
∂u ∂u
∂x ∂y
For u and v functions of two independent variables x and y, the determinant is
∂v ∂v
∂x ∂y
∂(u, v)
called* Jacobian of u, v with respect to x, y and is denoted by either ∂(x, y) or J(u, v).
Similarly Jacobian of u(x, y, –z) is given by
∂u ∂u ∂u
∂x ∂y ∂z
∂(u, v, w) ∂v ∂v ∂v
= J(u, v, w) =
∂(x, y, z ) ∂x ∂y ∂z
∂w ∂w ∂w
∂x ∂y ∂z
and this way it can be extended to number of variables.
∂(u, v) ∂( x, y)
I. Property: ⋅ = 1 or briefly J J' = 1.
∂(x , y) ∂(u, v)
Let u = u(x, y) and v = v(x, y) which on solving for x and y, gives, x = x(u, v) and y = y(u, v).
∂u ∂ u ∂ x ∂ u ∂y
=1= +
∂u ∂x ∂u ∂y ∂u
∂u ∂ u ∂ x ∂ u ∂y
=0= +
∂v ∂x ∂v ∂y ∂v
∂v ∂ v ∂ x ∂ v ∂ y
=0= +
∂u ∂x ∂u ∂y ∂u
∂v ∂ v ∂ x ∂ v ∂y
=1= +
∂v ∂ x ∂ v ∂ y ∂ v
∂u ∂ u ∂x ∂x ∂u ∂u ∂x ∂y
∂(u, v) ∂(x, y) ∂x ∂y ∂ u ∂ v = ∂x ∂y ∂u ∂u = 1
JJ' = ⋅ = ⋅
implying ∂(x, y) ∂(u, v) ∂v ∂ v ∂y ∂y ∂v ∂v ∂x ∂y
∂x ∂y ∂ u ∂v ∂x ∂y ∂v ∂v
(interchange rows and columns of 2nd determinant)
*Jacob-Jacobi (1804–1851) was a German Mathematician who made far reaching contributions in mechanics,
p. d. e., astronomy, elliptic functions and calculus of variations.
Partial Derivatives and their Applications 307
II. Chain Rule: If u, v are functions of r, s and r, s are further functions of x, y then
∂(u, v) ∂(u, v) ∂(r, s)
= ⋅
∂(x , y) ∂(r, s) ∂(x , y)
∂u ∂u ∂r ∂r ∂u ∂u ∂r ∂s
∂(u, v) ∂(r , s)
⋅ = ∂r ∂s ⋅ ∂x ∂y
= ∂r ∂s
∂x ∂x
∂(r, s) ∂(x, y) ∂ v ∂v ∂s ∂s ∂v ∂v ∂r ∂s
∂r ∂s ∂x ∂y ∂r ∂s ∂y ∂y
(on inter changing rows and columns in 2nd determinant)
∂u ∂r ∂u ∂s ∂u ∂r ∂u ∂s ∂u ∂u
+ +
∂r ∂x ∂s ∂x ∂r ∂y ∂s ∂y ∂x ∂y ∂(u, v)
= = =
∂v ∂r ∂v ∂s ∂ v ∂y ∂ v ∂ s ∂v ∂v ∂(x, y)
+ +
∂r ∂x ∂s ∂x ∂r ∂y ∂s ∂y ∂x ∂y
∂(u, v)
Observations: If u, v are functions of two independent variables x and y then u, v are independent of ≠0
∂(x, y)
(identically), otherwise, they are dependent and this property is extenable to system of any number of variables.
III. Jacobian of Implicit Functions: If u, v, w instead of being given explicitly in terms of x,
y, z be connected with them by equations such as
f1(u, v, w, x, y, z) = 0, f2(u, v, w, x, y, z) = 0, f3(u, v, w, x, y, z) = 0, then
∂( f1 , f2 , f3 )
∂(u, v, w) ∂(x, y, z)
= (−1)3 ,
∂(x, y, z) ∂ ( f1 , f2 , f3 )
∂(u, v, w)
dy f
Observations: This result can be generalized. It is in line to the result = − x , where x, y are connected by
dx fy
the relation f(x, y) = 0.
IV. Functional Relationship: If u, v, w be functions of x, y, z then the necessary and sufficient
condition for the existence of a functional relationship of the form f(u, v, w) = 0 is
u, v, w
J =0
x, y, z
Example 46:
∂( x , y )
θ, y = r sinθ
(i) In polar coordinates, x = r cosθ θ, show that = r.
∂(r , θ)
∂ ( x , y , z)
θ, y = r sinθ
(ii) In cylindrical coordinates, x = r cosθ θ, show that = r.
∂(r , θ, z)
θ cosφ
(iii) In spherical coordinates, x = r sinθ φ, y = r sinθ
θ sinφ
φ, z = r cosθ
θ, show that
∂ ( x , y , z)
= r2 sinθ. [NIT Kurukshetra, 2003]
∂(r , θ, φ)
Solution:
∂x ∂x
(i) For x = r cos θ, = cosθ, = −r sin θ
∂r ∂θ
308 Engineering Mathematics through Applications
∂y ∂y
y = r sin θ, = sinθ, = r cos θ
∂r ∂θ
∂x ∂x
∂(x, y) cos θ −r sin θ
= ∂r ∂θ = = r.
∴ ∂(r, θ) ∂y ∂y sin θ r cos θ
∂r ∂θ
∂x ∂x ∂x
(ii) For x = r cosθ, = cos θ, = −r sin θ, =0
∂r ∂θ ∂z
∂y ∂y ∂y
y = r sinθ, = sin θ, = r cos θ, =0
∂r ∂θ ∂z
∂z ∂z ∂z
z = z, = 0, = 0, =1
∂r ∂θ ∂z
∂x ∂x ∂x
∂ r ∂θ ∂ z cos θ −r sin θ 0
∂(x, y, z) ∂ y ∂ y ∂ y
∴ = = sin θ r cos θ 0 = r
∂(r, θ, z) ∂r ∂θ ∂z 0 0 1
∂z ∂z ∂z
∂ r ∂θ ∂ z
As z (3rd coordinates) remains –z, therefore, J becomes the same as in the case (i).
∂x ∂x ∂x
(iii) For x = r sinθ cosφ, = sin θ cos φ, = r cos θ cos φ, = −r sin θ sin φ;
∂r ∂θ ∂φ
∂y ∂y ∂y
y = r sinθ sinφ, = sin θ sin φ, = r cos θ sin φ, = r sin θ cos φ;
∂r ∂θ ∂φ
∂z ∂z ∂z
z = rcosθ, ∂r = cos θ, ∂θ = −r sin θ, ∂φ = 0;
∂x ∂x ∂x
∂r ∂θ ∂φ
sin θ cos φ r cos θ cos φ −r sin θ sin φ
∂(x, y, z) ∂ y ∂y ∂y
∴ = = sin θ sin φ r cos θ sin φ r sin θ cos φ
∂(r, θ, φ) ∂r ∂θ ∂φ cos θ −r sin θ 0
∂z ∂z ∂z
∂r ∂θ ∂φ
= r2[sinθcosφ (0 + sin2θcosφ) – cosθcosφ (0 – sinθcosθcosφ)
– sinθsinφ (–sin2θsinφ – cos2θ sinφ)]
= r2[sin3θ(cos2φ + sin2φ) – sinθcos2θ(cos2φ + sin2φ)]
= r2[sin3θ + sinθcos2θ] = r2sinθ(sin2θ + cos2θ) = r2sinθ
Example 47: If
x = u(1 − v)
y = uv }
, prove that J J’ = 1.
Solution: Given
x = u(1 − v)
y = uv }
gives
∂x
∂u
= (1 − v),
∂x
∂v
= −u
∂y ∂y
= v, =u
∂u ∂v
Partial Derivatives and their Applications 309
∂(x, y) 1 − v −u
so that J= = = u(1 − v) + uv = u. …(1)
∂(u, v) v u
Further x + y = u(1 − v) + uv = u u = (x + y)
implying y
and y = uv = (x + y) ⋅ v v=
x + y
∂u ∂v −y −uv
= 1;
u
= = 2 =− ;
∂x ∂x (x + y)2 u v
so that
∂u and ∂v x u(1 − v) 1 − v
=1 = = = .
∂y ∂y (x + y)2
u2 u
∂(u, v) 1 1
1− v v 1
J' = = −v 1− v = u + u = u
∂(x, y) …(2)
u u
1
∴ JJ' = − = 1. Hence the result.
u
x = a coshξ cosη ∂( x , y ) a
Example 48: If y = a sinhξ sinη , show that ∂(ξ , η) = 2 [ cos2ξ − cos2η ] [SVTU, 2007]
x = a cosh ξ cos η ∂x ∂x
Solution: gives ∂ξ = a sinh ξ ⋅ cos η, ∂η = −a cosh ξ ⋅ sin η
y = a sinh ξ sin η
∂y ∂y
= a cosh ξ ⋅ sin η, = a sinh ξ ⋅ cos η
∂ξ ∂η
∂x ∂x
∂(x, y) ∂ξ ∂η a sinh ξ ⋅ cos η −a cosh ξ ⋅ sin η
= =
∴ ∂(ξ, η) ∂y ∂y a cosh ξ ⋅ sin η a sinh ξ ⋅ cos η
∂ξ ∂η
(1 + cos 2η) (1 − cos 2η)
= a2 sinh2 ξ ⋅ cos2 η + cosh2 ξ ⋅ sin2 η = a2 sinh2 ξ + cosh2 ξ
2 2
a2
= (sinh2 ξ + cosh2 ξ) − (cosh2 ξ − sinh2 ξ)cos 2η
2
a2
= [cosh 2ξ − cos 2η] ; (using cosh2ξ + sinh2ξ = cosh2ξ, cosh2ξ – sinh2ξ = 1)
2
yz zx xy ∂(u, v, w)
Example 49: If u = x , v = y , w = z ; show that ∂(x , y, z) = 4.
yz zx xy
Solution: For given u = ,v = ,w = ,
x y z we know that
∂u ∂u ∂u −yz z y
∂x ∂y ∂z x2 x x
∂(u, v, w) ∂v ∂v ∂v z −zx x
= =
∂(x, y, z) ∂x ∂y ∂z y y2 y
∂w ∂w ∂w y x −xy
∂x ∂y ∂z z z z2
310 Engineering Mathematics through Applications
−yz zx xy −1 1 1
1 yz zx xy
= yz − zx xy = 1 −1 1 = 0 + 2 + 2 = 4.
x2 y2 z2 x2 y2 z2
yz zx xy 1 1 −1
x+y ∂(u, v)
Example 50: If u = , v = tan–1 x + tan–1y, then find ∂(x, y) . Are u and v functionally.
1 − xy
related? If so, find this relationship. [KUK, 2004, 2008]
x = tan θ θ = tan−1 x ∂θ 1 1
Solution: Take so that −1 and ∂x
= 2 = = cos2 θ
y = tan φ φ = tan y 1 + x 1 + tan 2
θ
…(1)
∂θ 1 1 2
= = = cos φ.
∂y 1 + y2 1 + tan2 φ
x+y tan θ + tan φ
Now u(x, y) = = = tan(θ + φ), …(2)
1 − xy 1 − tan θ tan2 φ
∂u ∂θ
giving = sec2 (θ + φ) = sec2 (θ + φ) ⋅ cos2 θ ,
∂x ∂x
∂u ∂θ …(3)
= sin2 (θ + φ) = sec2 (θ + φ)cos2 φ
∂y ∂y
Likewise, v(x, y) = tan–1x + tan–1y = (θ + φ) …(4)
∂v ∂v ∂θ
giving = = 1 ⋅ cos2 θ
∂x ∂θ ∂x
…(5)
∂v ∂v ∂θ
= = 1 ⋅ cos2 φ
∂y ∂θ ∂y
∂u ∂u
∴ ∂(u, v) ∂x ∂y sec2 (θ + φ)cos2 θ sec2 (θ + φ)cos2 φ cos2 θ cos2 φ
= = = sec2(θ + φ) = 0.
∂(x, y) ∂v ∂v cos2 θ cos2 φ cos2 θ cos2 φ
∂x ∂y
further from (2) and (4), u = tan v.
∂(x, y, z)
Example 51: If u = x + y + z, uv = y + z, uvw = z, show that = u2 v.
∂(u, v, w)
[VTU, 2003; NIT Kurukshetra, 2005; KUK 2009; PTU, 2009]
Solution: Here u = x + y + z x = u − (y + z) = u(1 − v) …(1)
uv = y + z ⇒ y = uv − z = uv(1 − w)
uvw = z z = uvw
∂x ∂x ∂x
∂u ∂v ∂w
∂(x, y, z) ∂y ∂y ∂y
Now =
∂(u, v, w) ∂u ∂v ∂w …(2)
∂z ∂z ∂z
∂u ∂v ∂w
Partial Derivatives and their Applications 311
∂x ∂y ∂z
From (1), = (1 − v), = v(1 − w), = vw,
∂u ∂u ∂u
∂x ∂y ∂z
= −u, ; = u(1 − w), ; = uw,
∂v ∂v ∂v
∂x ∂y ∂z …(3)
=0 = −uv = uv
∂w ∂w ∂w
On using (3), we get
−v −u 0
∂(x, y, z)
= v(1 − w) u(1 − w) −uv
∂(u, v, w)
vw uw uv
= (1 – v) {u(1 – w) uv + uv uw} + u{(1 – w) uv + uv uw}
= (1 – v) {u2v – u2vw + u2uw} + u{uv2 – uv2w + uv2w}
= (1 – v) u2v + u2v2
= u2v.
Alternately: Let x + y + z = l = u …(4)
y + z = m = uv
z = n = uvw
∂x ∂x ∂x
= 1, = −1, = 0
x =l−m ∂l ∂m ∂n
So that
∂y ∂y ∂y
y = m − n and = 0, = 1, = −1
z=n ∂l ∂m ∂n …(5)
∂z ∂z ∂z
= 0, = 0, =1
∂l ∂m ∂n
∂(x, y, z) ∂(x, y, z) ∂(l, m, n)
Now employ the property ∂(u, v, w) = ∂(l, m, n) ⋅ ∂(u, v, w) …(6)
∂x ∂x ∂x
∂l ∂m ∂n 1 −1 0
∂(x, y, z) ∂ y ∂y ∂y
whereas = = 0 1 −1 = 1
∂(l, m, n) ∂l ∂m ∂n 0 0 1 …(7)
∂z ∂z ∂z
∂l ∂m ∂n
∂l ∂l ∂l
∂u ∂v ∂w 1 0 0
∂(l, m, n) ∂m ∂m ∂m
= = v u 0 = u2 v
and ∂(u, v, w) ∂u ∂ v ∂w …(8)
∂n ∂n ∂n vw uw uv
∂u ∂v ∂w
Using (7) and (8), we get
∂(x, y, z)
= 1 ⋅ u2 v = u2 v.
∂(u, v, w)
312 Engineering Mathematics through Applications
ASSIGNMENT 6
∂(x, y)
1. If x = u2 – v2, y = 2uv; obtain .
∂(u, v)
∂(F, G, H )
2. If F = xu + v – y, G = u2 + vy + w, H = zu – v + vw; obtain .
∂(u, v, w)
∂(u, v)
3. If u = 2xy, v = x2 – y2; x = r cos θ, y = r sinθ; find .
∂(r, θ)
∂(u, v)
4. If u = sin– 1x + sin– 1y and v = x 1 − y2 + y 1 − x2 ; find . Are u and v fuctionally
∂(x, y)
related? If so, find this relationship. [KUK, 2006]
5. If u = xy + yz + zx, v = x2 + y2 + z2 and w = x + y + z, determine whether there is a
functional relationship between u, v and w. If so, find it.
6. Show that the functions u = x + y + z, v = x2 + y2 + z 2 – 2xy – 2zx – 2yz and
w = x3 + y3 + z3 – 3xyz are functionally related. Find the relationships between them.
[Hint: Follow Example 51 taking x + y + z = r, xy + yz + zx = s, x2 + y2 + z2 = t2]
∂(x, y, z)
7. If u = x + y + z, v = xy + yz + zx, w = xyz; evalutate ∂(u, v, w) . [HINT: JJ’ = – 1]
x y z ∂(u, v, w)
8. If u = ,v= ,w= , find the value of
∂(x, y, z) [NIT Kurukshetra, 2007]
y−z z−x x−y
4.8 ERRORS AND APPROXIMATIONS
Let u be a function of two independent variables x and y, and let δx and δy be small changes
made in x and y respectively resulting in a change in u by δu as
δu = (u + δu) – u = u(x + δx, y + δy) – u(x,y)
u(x + δx, y + δy) − u(x, y + δy) u(x, y + δy) − u(x, y)
= δx + δy
δx δy
∂u ∂u
= δx + δy approximately …(1)
∂x ∂y
u(x + δx, y + δy) − u(x, y + δy) ∂u u(x, y + δy) − u( x, y) ∂u
Since Lt = and Lt =
δx → 0 δx ∂x δy →0 δx ∂y
Well if δx and δy are taken as small errors in measurements of x and y, then δu given by (1)
above will be termed as an error in calculating value of u. This change in value of u by δu is
seen as from three point of view:
Absolute change: δu
δu
Relative change:
u
δu
Percentage change: × 100
u
In practice, percentage change (or error) and relative change (or error) are usually more
important than the absolute change (or error).
Partial Derivatives and their Applications 313
Further the expression (1) is valid for u being function of any number of variables say x, y,
z, t, … then,
∂u ∂u ∂u
δu = δx + δy + δz + … approximately …(2)
∂x ∂y ∂z
Note: For percentage error, it is always advisable to take log on both sides of the governing equation first and
then proceed as usually.
Example 52: How sensitive is the volume V = πr2h of a right circular cylinder to small
changes in its radius and height near the point (r0, h0) = (1, 3)?
Solution: By increment method, we get
∂V ∂V
∆V ≈ ∆r + ∆h
∂r (r , h ) ∂h (r , h )
0 0 0 0
= Vr(1, 3) ∆r + Vh(1, 3)∆h
= (2πrh)(1, 3) ∆r + (πr2)(1, 3) · ∆h
= 6π · ∆r + π · ∆h
The above result shows that a one-unit
change in r will change V nearly by 6 π units
and a one-unit change in h will change V nearly r=1
by 6π units. Therefore, the volume of a cylinder
with radius r = 1 and height h = 3 is nearly 6 r=3
h=3 h=1
times as sensitive to small change in r as it is to
small change in h Fig. 4.3 (i).
In contrast, if value of r and h are reversed,
so that r = 3 and h = 1, then ∆V ≈ 6π · ∆r + 9π · 1
4.3. (i ) h=
∆h. The volume is now more sensitive to small
4.3. (ii )
change in h to that it is to change in r. Thus the
Fig. 4.3
sensitivity to change depends not only on the
increment but also on the relative size of r and
h (See Fig. 4.3 (ii)).
Example 53: How sensitive is the change in V = πr2h related to the relative change in r and
h? How are the percentage changes related?
Solution: By error approximation, V = πr2h gives
∆V ≈ Vr · ∆r + Vh · ∆h = 2πrh · ∆r + πr2∆h …(1)
∆V 2πrh πr ∆r ∆h
2
or ≈ 2 ∆r + 2 ∆h = 2 + …(2)
V πr h πr h r h
Clearly the relative change in V is the relative change in h plus two times the relative
change in r.
∆V × 100 ≈ 2 ∆r × 100 + ∆h × 100
Further, V r h …(3)
indicates that the percentage change in V is the percentage change in h plus two times the
percentage change in r.
314 Engineering Mathematics through Applications
Example 54: If r is measured with an accuracy of ±1 percent and h with an accuracy of ±0.5
percent, about how accurately can we calculate V from the formula V = πr2h?
Solution: For V = πr2h or logV = log π + 2logr + log h, increment approximation implies
∆V ∆r ∆h
≈2 + . …(1)
V r h
∆r ∆r 1
× 100 = ±1, ≤ ,
r r 100
Given, ∆h 1 so that
∆h 1 …(2)
× 100 = ± 0.5 = ≤
h 2 h 200
∆V ∆r ∆h ∆r ∆h 2 1
∴ ≈ 2 + ≤2 + = + = 0.025 …(3)
V r h r h 100 200
∆V
Thus, the maximum percentage error (or change), viz. × 100, due to possible
V
percentage error (or change) in measurement of r and A will be about 2.5 per cent.
However, it is very difficult to answer such question for functions which involves two or
more independent variables because there could be several possible setting for errors in r
and h giving less than 2% error in V.
In such a case, we look for a reasonable square about the measured values (r0, h0) in which
V will not vary by more than the allowable value V0 = πr02ho.
Example 55: Find a reasonable square about the point (a, b) = (1, 3) in which the value of
V = πr2h will not vary more than ±0.1.
Solution: By change estimation,
∆V ≈ (2πrh)(r0, h0)∆r + (πr2)(r0, h0)∆h
= (2π × 1 × 3) ∆r + (π 9) ∆h = 6π ∆r + 9π ∆h …(1)
Since we have been looking for a square about the point (1, 3), we will take ∆h = ∆r so that
(1) becomes
∆V ¾ 6π ∆r + 9π ∆r = 15π ∆r …(2)
In order to get,
0.1
|∆V| ≈ 15π |∆r| ≤ 0.1 or ∆r ≤ ≈ 2.12 × 10−3 …(3)
15π
With ∆h = ∆r, the square we require about (1, 3) is given by
|r – 1| ≤ 2.12 × 10–3, |h – 3| ≤ 2.12 × 10–3
Hence as long as (r, h) lies within this square, we may expect |∆V| ≤ 0.1.
Example 56: The range R of a projectile which starts with a velocity v at an elevation α is
given by R = v sin2α . Find the percentage error in R due to an error of 1% in v and an
2
g
error of 0.5% in α. [KUK, 2009]
v2 sin2α
Solution: Given R = or logR = 2 logv + log sin 2α – log g.
g
Partial Derivatives and their Applications 315
By error approximation (i.e. taking differential on both sides)
δR δv 1
=2 + cos 2α ⋅ 2δα , (δg = 0)
R v sin 2α
or ( δRR × 100) = 2 ( δvv × 100) + 2α(cot 2α)( δαα × 100)
1
= 2 × 1 + 2α ⋅ cot 2α × 0.5 = 2 + 2α cot 2α ×
2
= (2 + α cot 2α)
Hence, the percentage error in calculation of R due to errors of 1% in R and 0.5% in α is
(2 + α cot 2α).
2
Example 57: If the Kinetic energy K = wv , find approximately the change in the kinetic
2g
energy as w changes 49 to 49.5 and v changes from 1600 to 1590.
wv2
Solution: Here K = so that δK = 1 [v2δw + w2vδv] approx.
2g 2g
On putting,
v = 1600 f.p.s. units, w = 49 f.p.s units, δv = –10 f.p.s. units, δw = 0.5 f.p.s. units, g = 32.
1
δK = 0.5(1600)2 + 2 × 49 × 1600 × (−10) = −4500 units ( f .p.s.)
2 × 32
Example 58: Find the percentage error in the area of an ellipse. When an error of 1% is
made in measuring the major and minor axes.
Solution: Let x and y be the major and minor axes of the ellipse, then area A = πxy implying
logA = logπ + logx + logy.
By approximation,
δA δx δy δA × 100 = δx × 100 + δy × 100 = 1 + 1 = 2.
= +
y
or
A x y A x
Hence the percentage error in measurement of the area of the ellipse with the percentage
error of 1% each in measurement of major and minor axes is 2 percent.
Example 59: Find the possible percentage error in computing the resistance r from the
1 1 1
formula = + if r1, r2 are both in error by 2%. [UPTech, 2004]
r r1 r2
1 1 1
Solution: By error approximation, r = r + r gives
1 2
1 1
d ≈ d +
1 1 1 1
r
or − δr ≈ − 2 δr1 − 2 δr2
r1 r2 r2 r1 r2
316 Engineering Mathematics through Applications
1 δr 1 δr 1 δr
implying 1 δr 1 δr1 1 δr2
; + or × 100 ≈ 1 × 100 + 2 × 100
r r r1 r1 r2 r2 r r r1 r1 r2 r2
δr × 100 ≈ r 1 ⋅ 2 + 1 ⋅ 2 = 2r 1 + 1 = 2, δr1 δr2
r r1 r2 r1 r2 as r × 100 = r × 100 = 2
1 2
Hence in the percentage error in r is 2 percent when r1 and r2 has possible percentage error
of 2 percent in each.
Example 60: The height h and the semi-vertical angle α of a cone are measured and from
them ‘A’, the total area of the cone, including the base is calculated.
If h and α are in error by small quantities δh and δα respectively, find the corresponding
π
error in the area. Show further that if α = , an error of 1% in h will approximately
6
compensated by an error of –0.33° in α. A
Solution: For a cone of height h, semi-vertical angle α,
radius of the base r and curved length l, we get
Area of the base = πr2 = π(h tanα)2 = πh2tan2α.
l = h sec α
Area of the curved surface = πrl = π(htanα)(h sec α)
= πh2 tanα secα h
∴ Total area, A (h,α) = πh2 tan2α + πh2 tanα secα
= πh2(tan2α + tanα secα)
∂A ∂A B r C
Implying δA = δh + δα approximately r = h tanα D
δh δα
= 2πh(tan2α + tan α secα)δh Fig. 4.4
+ πh2(2tanα · sec2α + tan2α · secα)δα
which gives the error in area, A(h, α) corresponding to errors δh and δα in h and α respectively.
π δh h
Now taking α= and × 100 = 1 or δh = , we get
6 h 100
2πh2 1 2 1 2 1 2 2 1 2 2
δA =
3 + + πh2
2 3 ⋅ 3 + 3 3 δα
100 3 3
2
If δA = 0 then 0 = π h2 (1) + πh2 (3.4646) ⋅ δα
100
Since π and h are non zero, therefore,
.02 .02
δα = − radians = − 57.3° = − 0.33°.
3.4646 3.4646
8π I l
Example 61: The torsional rigidity of a length of a wire is from the formula N = . If l
t2 r 4
is decreased by 2%, r is increased by 2%, t is increased by 1.5%, show that the value of N
is diminished by 13% approximately.
Partial Derivatives and their Applications 317
Solution: Given,
δl
l is decreased by 2% i.e.× 100 = −2
l
δr
r is increased by 2% i.e. × 100 = 2
r
δt 3
t is increased by 1.5% i.e. × 100 = 1.5 =
t 2
8πI l
Torsional rigidity, N(l, t, r) =
t2 r4
implies log N = log(8πI) + logl – 2logt – 4logr …(1)
Taking differentials on both sides,
δN δ l δt δr
= −2 −4
N l t r
or
δN × 100 = δ l × 100 − 2 δ t × 100
N l t (δr
−4 r × 100 )
3
= −2 − 2 ×
− 4 × 2 = −13
2
Hence N diminishes by 13% if with the above given percentage changes in l, r, t.
Example 62: At a distance of 50 metres from the foot of the tower the elevation of its top is
30°. If the possible errors in measuring the distance and elevation are 2 cm and 0.05 degrees,
find the approximate error in calculating the height. [NITK, 2002; UPTech, 2004]
A(top)
Solution: The given problem with h as height of the top
(point A) from the bottom (point B) and α the elevation of
the top with the ground is α° as explained in Fig. 4.5. From h = x tanα
the figure, it is apparant that height h is a function of the
elevation, α and the distance of point of elevation from
α
bottom, x, i.e. C B(bottom)
x = 50 cm
h(x, α) = x tanα …(1)
Fig. 4.5
so that δh = hxδx + hαδα = tan α δx + x sec2α δα …(2)
2 5 π
For given x = 50, δx = 2 cm = mt; α = 30°, δα = 0.05° = radians,
100 100 180
2
2 5 π 1 2 2 5 π
δh = (tan 30 °) ⋅ + 50 ⋅ sec2 30° ⋅ = + 50 ⋅
100 100 180 3 100 3 100 180
= 0.0116 + 0.0582 = 0.0698 ¾ 0.07 mts. = 7 cms.
Example 63: If the sides and angles of a triangle ABC vary in such a way that its circum-
da db dc
radius remains constant. Prove that + + =0
cos A cos B cos C
318 Engineering Mathematics through Applications
Solution: The circum-radius of a triangle ABC is given by
a b c
R= = =
2 sin A 2 sin B 2 sin C
or a = 2R sin A, where R is constant …(1)
= 2R dA … (i)
da
implying da = 2R cosA dA or
cos A
db
= 2R dB … (ii)
cos B
dc
= 2R dC … (iii)
cos C
Adding (i), (ii), (iii),
da db dc
+ + = 2R (dA + dB + dC) …(2)
cos A cos B cos C
Also in triangle ABC,
∠ A + ∠ B + ∠ C = π (an implicit relation) or dA + dB + dC = 0 …(3)
Thus on using (3) in (2), we get
da db dc
+ + =0
cosA cosB cosC
Problem on Relative Measurement
1
Example 64: What will be error in calculating the area of a triangle, ∆ = bc sinA , if the
2
two sides have an error of ± 1 % each and error in A is 1 minute of the angle in measurement
4
for A = 60°?
1 1
Solution: Here ∆= bc sin A implies log ∆ = log 2 + log b + log c + log sin A …(1)
2
Taking differentials on both sides,
∆ b c 1
= + + cos A δ A
δ∆ δ b δ c sin A
∆ × 100 = b × 100 + c × 100 + cot A × δ A × 100
or …(2)
δ∆ δb δc
b 1 c 1
Given, × 100 = , × 100 = ; …(3)
δb 4 δc 4
δA π
or δA = A =
1 1 1
and = 1' = × 60 = 1 ° = radians …(4)
A 60 ° 60 60 180
Partial Derivatives and their Applications 319
∆ 1 1 π 1 1 22 1
Using (3, 4), × 100 = + + cot 60 ° ⋅ × 100 = + × × 100
δ∆ 4 4 180 2 3 7 180
= 0.5 + (1.74603)
1
1.732050
= 0.5 + 1.00807 = 1.50807
Miscellaneous Problems
Example 65: How accurately should length (l) and time of vibration (T) of a pendulum
should be measured in order that the computed value of g be correct to 0.01%.
Solution: We know that for a pendulum,
l l
T = 2π or g = 4π2 …(1)
g T2
Means g is a function of l, the length of the pendulum and T, the time period of oscillations.
∂ g 4π2 ∂g 8π2l
From (1), = 2 and =− 3 …(2)
∂l T ∂T T
We know that, inversely to find errors in x1, x2, …, xn when X = f(x1, x2, …, xn) is to have
a desired accuracy
i.e. if ∆X is the error in X, we have to determine errors ∆x1, ∆x2, …, ∆xn in x1, x2, … , xn as
∂X ∂X ∂X
∆X = ∆ x1 + ∆ x2 + … + ∆ xn …(3)
∂x1 ∂x2 ∂xn
However, on using the principal of equal effects, viz.
∂X ∂X ∂X
∆x1 = ∆x2 = … = ∆xn …(4)
∂x1 ∂x2 ∂xn
equation (3) results in
1 ∆X 1 ∆X 1 ∆X … (5)
∆x1 = , ∆x2 = , ∆xn =
n ∂X n ∂X n ∂ X
∂x1 ∂x2 ∂xn
On using above concept, we get
1 ∆g 1 ∆g
∆l = and ∆T =
∂
2 g ∂
2 g
…(6)
∂l keeping T constant ∂T keeping l constant
320 Engineering Mathematics through Applications
So that percentage error in l viz.
∆l 1 1 ∆g 1 ∆g
× 100 = × 100 = × 100 , (using (2))
l l 2 ∂g 2l 4π2
2
∂l T
1 ∆g 1 ∆g 1
= 2 × 100 = × 100 = × 0.01 = 0.005
2 4π l 2 g 2 …(7)
T2
and percentage error in T viz.
∆T 1 1 ∆g 1 ∆g
× 100 = × 100 = × 100
T T 2 ∂g 2T − 8π2l
3
∂T T
− 1 ∆g − 1 ∆g −1
= 2 × 100 = × 100 = × 0.01 = 0.0025
4 4π l 4 g 4 …(8)
2
T
Hence in order to have a percentage error in g by 0.01%, percentage errors in l and T must
not exceed by 0.005 and 0.0025, respectively.
r2 h
Example 66: The percentage error in R which is given by R = + , is not allowed in r
2h 2
and h when r = 4.5 cm and h = 5.5 cm.
∆R
Solution: The percentage error in R = × 100 = 0.2 …(1)
R
0.2 0.2 r2 h 0.2 (4.5)2 5.5
∴ ∆R = R= + = +
100 100 2h 2 100 2 × 5.5 2
0.2 50.50 0.002 × 50.50
= × = …(2)
100 11 11
∆r
Now, percentage error in r = × 100
r
100 ∆R 100 h ∆R
= =
r 2 ( r h) 2r2
100 0.002 × 50.50
= × 5.5 × = 0.12
2(4.5)2 11
Partial Derivatives and their Applications 321
∆h
Percentage error in h = × 100
h
100 ∆R
=
h r2 1
2− 2 +
2h 2
100 ∆ R
=
r2
− + h
2h
100 50.50 × .002
= × = 5.05
20 11
11
ASSIGNMENT 7
1. Find the percentage error in the area of a rectangle when an error of 1% is made in
measuring its length and breadth.
2. The time oscillations of a simple pendulum is given by T = 2π l . In an experiment
g
carried out to find the value of g, errors of 1% and 0.5% are possible in the values of l
and T respectively. Show that the maximum errors in the calculated value of g could not
exceed 2%. [NIT Kurukshetra, 2008]
[Hint: Possible pairs of errors (±1, ±0.5)]
3. If Pv2 = k and the relative errors in p and v are respectively 0.50 and 0.025, show that the
error in k is 10%.
4. If the H.P. required to propel a steamer varies as the cube of the velocity and square of
the length. Prove that a 30% increase in velocity and 4% increase in length will require
an increase of about 17% in H.P.
5. The deflection at the centre of a rod of length l and diameter d supported at its ends,
located at the centre with a weight w varies as wl3d– 4. What is the increase in deflection
corresponding to p% increase in w, q% increase in l and r% increase in d ?.
6. The work must be done to propel a ship of displacement D for a distance s in time t is
2
proportional to
S2D 3
. Find approximately the increase of work necessary when the
t2
displacement is increased by 1%, the time is diminished by 1% and the distance is
diminished by 2%.
7. The indicated horse power l of an engine is calculated from the formula
l = PLAN/33,000, where A = π d . Assuming that the error of r percernt may have been
2
4
made in measuring P, L, N and d, find the greatest possible error in l.
322 Engineering Mathematics through Applications
8. The length of the side c of a triangle is calculated from the rest two sides a, b and the
included angle C. Show that the total error in c due to errors δa, δb, δC in a, b, C is given
by δacosB + δbcosA + aδCsinB
[Hint: Take c2 = a2 + b2 – 2ab cos C, Use b cos C + c cos B = a]
9. Two sides a, b of a triangle and including angle C are measured. Show that the error δc
in the computed length of the 3rd side c due to small errors in the angle C is given by
aδC sin B ·
a2 + b2 − c2
Hint : Use cos C = with a, b both constant
2ab
10. If ∆ be the area of a triangle, prove that the error in ∆ resulting from small error in c is
given by
1 −1
δ∆ = ∆[s + (s − a)−1 + (s − b)−1 − (s − c)−1 ] δ c.
4
Hint : ∆ = s(s − a)(s − b)(s − c) where s = a + b + c
2
11. In a plane triangle ABC if the sides a and b be kept constant, show that the variation of
its angle are given by the relation
dA dB dC
= =
a2 − b2 sin2 A b2 − a2 sin2 B −c
Hint : Use a = b , a cos B + b cos A = c
sin A sin B
12. Show that the acceleration due to gravity is reduced by nearly 1% at an altitude equal to
0.5% of earth’s radius, given that at an external point x kilometres from earth’s centre
2
such an acceleration is given by g , where r is the radius of the earth.
r
x
[Hint: Here a = g(r/x)2, with g and r constant]
4.9 TAYLORS THEOREM FOR FUNCTIONS OF TWO VARIABLES
Statement: Suppose we have a function of two variables, z = f (x, y) which is continuous
together with all its partial derivatives upto (n + 1)th order inclusive, in some neighbourhood
of the point (a, b), then like in case of one variable, z = f(x, y) is expressible as:
∂ ∂
f (a + h, b + k) = f (a, b) + h + k f (a, b)
∂x ∂y
2 n
1 ∂ ∂ 1 ∂ ∂
+ h + k f (a, b) + … + h + k f (a, b) + Rn …(1)
2 ∂x ∂y n ∂x ∂y
where Rn is the remainder term given as:
Partial Derivatives and their Applications 323
n+1
1 ∂ ∂
Rn = h +k f (a + θh, b + θk), 0 < θ < 1 …(2)
n + 1 ∂x ∂y
If this remainder Rn → 0 as h → ∞, then the Taylor’s Theorem becomes Taylor’s Infinite
series as:
2
∂ ∂ 1 ∂ ∂
f (a + h, b + k) = f (a, b) + h + k f (a, b) + h + k f (a, b) + ……∞ …(3)
∂x ∂y 2 ∂x ∂y
Proof: Let x = a + ht, y = b + kt, where t is the parameter which takes the values in the interval
[0, 1]. Define a function
F(t) = f(x, y) = f(a + ht, b + kt) …(4)
Then by chain rule on (1),
d ∂ f dx ∂ f dy ∂ f ∂f
F(t) = F'(t) = + = h +k
dt ∂ x dt ∂ y dt ∂ x ∂y
Precisely, ∂ ∂
F'(t) = h +k f,
∂x ∂y
2
∂ ∂
F"(t) = h + k f,
∂x ∂y
....................................
....................................
n+1
…(5)
n +1 ∂ ∂
F (t) = h +k f
∂x ∂y
On using Taylor’s Theorem for functions of one variable, when t = 1, a = 0, we obtain
1 n 1
F(1) = F(0) + F'(0) + … + F (0) + Fn+ 1(0) …(6)
n n+1
where F(1) = f (a + h, b + k )
F(0) = f (a, b)
∂ ∂
F'(0) = f h + k f (a, b)
∂x ∂y
2
∂ ∂
F"(0) = h + k f (a, b)
x ∂ ∂ y …(7)
………………………………
………………………………
n+1
∂ ∂
Fn +1(0) = h +k f (a + θh, b + θk ), 0 < θ < 1
∂x ∂y
With above values of F(1), F(0), F’(0), …, F (n + 1) (0), in (6), we get the Taylor’s Theorem for
functions of two variables given by equations (1) and (2) in the statement.
324 Engineering Mathematics through Applications
On substituting x = a + h, y = b + k, we can also write Taylor’s Theorem as:
2
∂ ∂ ∂ ∂
f (x, y) = f (a, b) + (x − a) + (y − b) f (a, b) + (x − a) + (y − b) f (a, b)
∂x ∂y ∂x ∂y
n
∂ ∂
+…… + (x − a) + (y − b) f (a, b) + Rn …(1)
∂x ∂y
n +1
1 ∂ ∂
where Rn = (x − a) + (y − b) f (ξ, η) …(2)
n + 1 ∂x ∂y
and ξ = (1 – θ)a, η = (1 – θ)b, 0 < θ < 1.
Further, if Rn → 0 as n → ∞, Taylor’s theorem becomes Taylor’s infinite series as below:
∂ ∂
f (a + h, b + k) = f (a, b) + (x − a) + (y − b) f (a, b)
∂ x ∂ y
2
1 ∂ ∂
+ ( x − a) + (y − b) f (a, b) + ……∞
2 ∂x ∂y …(3)
Note: 1. Expansion of f(x, y) as given in (3) is known as Taylor’s expansion of f(x, y) about the point (a, b).
Expansion of f(x, y) as given in (3) is also known as Taylor’s expansion of f(x, y) in powers of (x
– a) and (y – b).
2. If we put x = 0, y = 0 and replace h = x, k = y in (3), we arrive at the extension of Maclaurin’s Theorem
for two independent variables.
Example 67: Expand (x2y + 3y – 2) in powers of (x – 1) and (y + 2) using Taylor’s Theorem.
Solution: Expansion
f (x, y) = f (a, b) + (x − a) fx (a, b) + (y − b) fy (a, b)
1
+
[(x − a)2 fxx( a, b) +2(x − a)(y − b) fxy (a, b) + (y − b)2 fyy (a, b)] + …
2
is Taylor’s expansion of f(x, y) in powers of (x – a) and (y – b).
whereas in this case, given x − a = x − 1, so that, a = 1,
y − b = y + 2 b = −2
Here
f(x, y) = x2y + 3y – 2, f(1, –2) = –10;
fx(x, y) = 2xy, fx(1, –2) = –4;
fy(x, y) = x2 + 3, fy(1, –2) = 4;
fxx(x, y) = 2y, fxx(1, –2) = –4;
fxy(x, y) = 2x, fxy(1, –2) = 2;
fyy(x, y) = 0, fyy(1, –2) = 0;
Partial Derivatives and their Applications 325
fxxx(x, y) = 0, fxxx(1, –2) = 0;
fxxy(x, y) = 2, fxxy(1, –2) = 2;
fxyy(x, y) = 0, fxyy(1, –2) = 0;
fyyy(x, y) = 0, fyyy(1, –2) = 0;
1
∴ x2 y + 3y − 2 = −10 + [(x − 1) − 4 + (y + 2)4] + (x − 1)2 (−4) + 2(x − 1)(y + 2)2 + (y + 2)2 ⋅ 0
2
1
+[(x − 1)3 ⋅ 0 + 3(x − 1)2(y + 2) ⋅ 2 + 3(x − 1)(y + 2)2 0 + (y + 2)3 ⋅ 0]
3
+ next all terms become zero
= – 10 – 4(x – 1) + 4(y + 2) – 2(x – 1)2 + 2(x – 1)(y + 2) + (x – 1)2(y + 2)
Example 68: Expand f(x, y) = tan–1 (y/x) in powers of (x – 1) and (y – 1) upto 3rd degree
terms. Hence Compute F(1.1, 0.9) approximately.
Solution: Taylor’s Expansion of f(x, y) in terms of (x – a) and (y – b) (or about (a, b)) is given
as:
f(x, y) = f(a + b, b + c) = f(a, b) + [(x – a)fx + (y – b)fy]
1
+ [(x − a)2 fxx + 2(x − a)(y − b) fxy + (y − b)2 fyy ] + …
2
x − a = x − 1
but here y − b = y − 1 so that (a, b) = (1, 1)
y
f (x, y) = tan−1 , f(1, 1) = tan–1 = π/4;
x
−y 1
fx (x, y) = , fx (1,1) = − ;
x + y2
2
2
x 1
fy (x, y) = , fy (1,1) = ;
x + y2
2
2
2xy 1
fxx (x, y) = , fxx (1,1) = ;
(x2 + y2 )2 2
y2 − x2
fxy (x, y) = , fxy(1, 1) = 0;
(x2 + y2 )2
−2xy 1
fyy (x, y) = , fyy (1,1) = − ;
(x + y2 )2
2
2
2y3 − 6x2 y 1
fxxx (x , y) = , fxxx (1,1) = − ;
(x2 + y2 )3 2
326 Engineering Mathematics through Applications
2x3 − 6xy2 1
fxxy (x , y) = , fxxy (1,1) = − ;
(x2 + y2 )3 2
6x2 y − 2y3 1
fxyy (x , y) = , fxyy (1,1) = ;
(x2 + y2 )3 2
6xy2 − 2x3 1
fyyy (x , y) = , fyyy (1,1) = ;
(x2 + y2 )3 2
∴ f (x, y) = tan −1(y / x) = f (1, 1) + (x − 1) fx (1, 1) + (y − 1) fy (1, 1)
1
+ (x − 1)2 fxx (1, 1) + 2(x − 1)(y − 1) fx (1, 1) + (y − 1)2 fyy (1, 1)
2
1
+ [(x − 1)3 fxxx (1, 1) + 3(x − 1)2(y − 1) fxxy (1,1) + 3(x − 1)(y − 1)2 fxyy + (y − 1)3 fyyy ] + …
3
π 1 1 1
+ (x − 1) − + (y − 1) + (x − 1)2 + 2(x − 1)(y − 1) ⋅ 0 + (y − 1)2 −
1 1
=
4 2 2 2 2 2
1 1
(x − 1)3 − + 3(x − 1)2(y − 1) − + 3(x − 1)(y − 1)2 + (y − 1)3 +…
1 1 1
+
3 2 2 2 2
π 1 1
= − (x − 1) − (y − 1) + (x − 1)2 − (y − 1)2
4 2 4
1
(x − 1)3 + 3(x − 1)2(y − 1) −3(x − 1)(y − 1)2 − (y − 1)3
−
12
For, f(1.1, 0.9) which is comparable to f (a + h, b + k),
take 1.1 = 1 + 0.1 so that a = 1, h = 0.1;
0.9 = 1 – 0.1 b = 1, k = – 0.1
π 1
4 2[
− 0.1 − (−0.1)] + (0.1)2 − (−0.1)2
1
∴ f (x, y) = f (1.1, 0.9) =
4
1
(0.1)3 + 3(0.1)2(−0.1) − 3(0.1)(−0.1)2 − (−0.1)3 + …
+
12
= 0.7854 – 0.0967 = 0.6887.
Example 69: Expand exsiny in powers of x and y as far as terms of 3rd degree.
Solution: Expansion of ex sin y in powers of x and y means it is a Maclaurin’s expansion viz.
expansion about (0, 0).
1
Write f (x, y) = f (a, b) + h fx(a, b) + k fy(a, b) + h fxx(a, b) + 2hk fxy(a, b) + k2 fyy(a, b) +……∞
2
when a = 0 = b and h = x, k = y
1 2
So that f (x, y) = f (0, 0) + h fx(0, 0) + k fy(0, 0) + h f (0, 0) + 2hk fxy(0, 0) + k2 fyy(0, 0) + ……∞
2 xx
Partial Derivatives and their Applications 327
Here f(x, y) = ex siny, f(0, 0) = 0
fx(x, y) = ex siny, fx(0, 0) = 0
fy(x, y) = ex cosy, fy(0, 0) = 1
fxx(x, y) = ex siny, fxx(0, 0) = 0
fxy(x, y) = ex cosy, fxy(0, 0) = 1
fyy(x, y) = – ex siny, fyy(0, 0) = 0
fxxx(x, y) = ex siny, fxxx(0, 0) = 0
fxxy(x, y) = ex cosy, fxxy(0, 0) = 1
fxyy(x, y) = –ex siny, fxyy(0, 0) = 0
fyyy(x, y) = –ex cosy, fyyy(0, 0) = –1
1 2
∴ f (x, y) = ex sin y = 0 + x ⋅ 0 + y ⋅ 1 + x ⋅ 0 + 2xy ⋅ 1 + y2 ⋅ 0
2
1
+ x3 ⋅ 0 + 3x2 y ⋅ 1 + 3xy2 ⋅ 0 + y3 ⋅ 1 + …
3
x2 y y3
= y + xy + − (upto 3rd degree)
2 6
1 1
Example 70: Evaluate log (1.03)3 + (0.98)4 − 1 approximately
1 1
1 1
Solution: Let f (x, y) = log x 3 + y 4 − 1 which is comparable to log (1.03)3 + (0.98)4 − 1 .
Now f(x, y) = f(a + h, b + k) = f(a, b) + [h fx(a, b) + k fy(a, b)]
1 2
+ h fxx(a, b) + 2hk fxy(a, b) + k2 fyy(a, b) + …
2
Here in the problem,
x = a + h = 1.03 = 1 + 0.03 so that (a, b) = (1, 1)
y = b + k = 0.98 = 1 − 0.02 (h, k) = (0.03, –0.02)
1 1
f (x, y) = log x 3 + y 3 − 1 , f(1, 1) = log[1 + 1 – 1] = log 1 = 0;
−2
1 3 1
x
fx (x, y) = 3 , 3 1
1 1 fx (1,1) = = ;
x3 + y4 − 1 1+1−1 3
−3
1 4 1
y
fy (x, y) = 4 , 4 1
1 1 fy (1,1) = = ;
x3 + y 4 − 1 1+1−1 4
328 Engineering Mathematics through Applications
1 1
∴ f (x, y) = log (1.03)3 + (0.98) 4 − 1 ; f (1, 1) + 0.03 fx (1, 1) + (−0.02) fy (1, 1) + …
1
= 0 + 0.03 − 0.02 = 0.005 approx.
1
3 4
ASSIGNMENT 8
π
1. Expand excosy about the point 1, 4 by Taylor’s Theorem.
π
2. Expand f(x, y) = sin xy in powers of (x – 1) and y − 2 upto the second degree term.
3. If f(x, y) = tan–1xy, compute f(0.9, –1.2) approximately.
4. Expand sin x sin y in powers of x and y as far as terms of third degree.
5. Expand ex·log(1 + y) in powers of x and y upto terms of third degree. [PTU, 2009]
4.10 MAXIMA-MINIMA OF TWO FUNCTIONS
Definition: A function f(x, y) of two variables is said to be maximum at (a, b) if f(a + h, b +
k) – f(a, b) < 0 for sufficiently small positive or negative values of h and k, and minimum if
f(a + h, b + k) – f(a, b) > 0.
In other words, if f(a + h, b + k) – f(a, b) = ∆ is negative for small values of h, k, then f(a, b)
is a maximum and if f(a + h, b + k) – f(a, b) = ∆ is positive for small values of h, k, then f(a, b) is
minimum.
The points at which maximum or minimum values occur are also known as points of
extrema of critical points and the maximum and minimum values taken together are extreme
values of the function.
Observations:
1. A function f(x, y) may also attain its extreme values on the boundary.
2. The maxima-minima so defined are local relative maxima or local relative minima. Thus, a maximum
value may not be the greatest and minimum may not be the least of all the values of the function in any
finite region.
3. The greatest and smallest values attained by a function over the entire region including the boundary
are called global (absolute) maximum and global (absolute) minimum values of the function.
E.g. for a function z = f(x, y) say representing a dom, maximum value of z occurs at the top from where
surface descends in all directiosns. If z = f(x, y) represents the equation of a bowl, minimum is attained
at the bottom from where surface ascends in all directions. Otherwise, a maximum or minimum value
may form a ridge such that the surface ascends or descends in all directions.
Besides, there are points on the surfaces, from where surface rises for displacement in certain
directions and fall for displacement in the other directions, called saddle points.
Necessary Conditions for Having Extremum: fx(a, b) = 0 = fy(a, b).
By Taylor’s Theorem,
( )
f (a + h, b + k) = f (a, b) + h fx (a, b) + k fy(a, b) +
1 2
2
( )
h fxx (a, b) + 2hkfxy (a, b) + k 2 fyy (a, b) + … …(1)
Partial Derivatives and their Applications 329
As h and k are very small, neglecting second and higher order terms, we get
∆ = h fx(a, b) + k fy(a, b). …(2)
From above, sign of ∆ depends on the sign of h fx(a, b) + k fy(a, b) which is a function of h
and k. Letting h → 0, we observe that ∆ changes sign when k changes sign. Therefore, the
function cannot extreme unless fy = 0.
Similarly, letting k → 0, we see that f(x, y) cannot have extremum unless fx = 0.
Hence the necessary condition for f(x, y) to have a maximum or minimum at (a, b) is
fx(a, b) = 0 and fy(a, b) = 0
If fx(a, b) = 0 = fy(a, b), then f(a, b) is called the stationary value of f(x, y) at (a, b).
Sufficient Conditions for a Function to have Extremum:
When necessary conditions are satisfied, then for small values of h and k, (1) reveals
1 2
sign of ∆ = sign of h r + 2hks + k2t , where r = fxx(a, b), s = fxy(a, b), t = fyy(a, b).
2
1 2 2
= sign of h r + 2hkrs + k2rt
2r
1
= sign of (h2r2 + 2hkrs + k2 s2 ) − k 2 s2 + k2rt
2r
1
= sign of (hr + ks)2 + k 2(rt − s2 )
2r
…(3)
In (3), (hr + ks)2 is always positive and k2(rt – s2) can be made positive if (rt – s2) > 0.
Thus with the condition that (rt – s2) > 0, ∆ changes sign with the change in sign of r.
Hence, if (rt – s2) > 0, f(x, y) has a Maximum or Minimum according as r < 0 or r > 0.
If (rt – s2) < 0, then sign of ∆ in (3) depends on sign of h and k, and hence no maximum or
minimum of f can occur at (a, b), i.e. it is a saddle point.
If, rt – s2 = 0 or r = t = s = 0, no conclusion can be made and further investigation is required
under such circumstances.
Observations: The sufficient condition for a critical point (a, b) may be taken as:
point of maximum if (rt – s2) > 0 and t < 0
point of minimum if (rt – s2) > 0 and t < 0
Since whenever extremum exists, then (rt – s2) > 0 and both r and t have the same sign either positive or
negative.
Example 71: Find the extreemum values of x3 + y3 – 3axy
OR
Determine the points where the function (x + y3 – 3axy) has a maximum or minimum.
3
Solution: The essential condition for Maxima-Minima is fx(x, y) = 0 = fy(x, y) or points of
Maxima-Minima are given by
∂f
= 3x2 − 3ay = 0 …(1)
∂x
330 Engineering Mathematics through Applications
∂f
and = 3y2 − 3ax = 0 …(2)
∂y
x2
From (1), y = , on substituting into (2),
a
x – a x = 0 or x(x – a)(x2 + ax + x2) = 0
4 3
giving either x = 0 or x = a (on neglecting imaginary roots).
x2 x2
So that y= = 0 for x = 0 and y = = a for x = a.
a a
Finding, the stationary points are at (0, 0) and (a, a).
∂2 f ∂2 f ∂2 f
r= = 6 x, s = = −3 a, t = = 6y
∂x2 ∂ x∂ y ∂y2
At (0, 0), (rt – s2) = 6x · 6y – 9a2 = 36xy – 9a2 = – 9a2 < 0 (for all values of a), while r = 0.
Hence it is a saddle point.
At (a, a), (rt – s2) = 36xy – 9a2 = 36a2 – 9a2 = – 27a2 > 0 (for all values of a) and r = 6a.
Hence point (a, a) gives a maximum if a is negative and minimum if a is positive, for r > 0.
Example 72: Examine the function f(x, y) = x4 + y4 – 2x2 + 4xy – 2y2 for extreme values.
Solution: For f (x, y) = x4 + y4 – 2x2 + 4xy – 2y2,
∂f ∂f
= 4x3 − 4x + 4y, = 4y3 + 4x − 4y
∂x ∂y
r = fxx = 12x2 – 4, s = fxy = 4, t = fyy = 12y2 – 4.
For finding extreme values of f(x, y), necessary conditions is fx = 0 = fy.
means 4x3 – 4x + 4y = 0 …(1)
4y3 + 4x – 4y = 0 …(2)
Adding (1) and (2),
4(x3 + y3) = 0 or (y + x)(y2 – xy + x2) = 0 or y = – x.
so that y = −x = 0, − 2, 2
Hence the possible points for extreme values are (0, 0), ( )(
2, − 2 − 2, )
2 .
At (0, 0), (rt − s2 ) = (12x2 − 4)(12y2 − 4) − 16 = 0 and hence further investigation is
(0,0)
needed.
Otherwise also, f (x, y)
(0,0)
= x4 + y4 − 2x2 + 4xy − 2y2 = 0 and for points along x-axis, where
y = 0, i.e. f (x, y) x,0 = x4 − 2x2 = x2(x2 − 2) .
Means in the neighbourhood of the origin there are points where f(x, y) has values greater
and lesser value than the value at (0, 0). Thus f(x, y) is neither maximum nor minimum at
(0, 0).
Partial Derivatives and their Applications 331
Example 73: Locate the stationary points of x2y2 – 5x2 – 8xy – 5y2. Discuss their nature.
Solution: For f(x, y) = x 2 y 2 – 5x 2 – 8xy – 5y 2 , stationary points are given by
fx(x, y) = 0 = fy(x, y) i.e.
∂f
= 2xy2 − 10x − 8y = 0 …(1)
∂x
∂f
= 2x2 y − 8x − 10y = 0 …(2)
∂y
On substracting (2) from (1),
2xy(y – x) + 2(y – x) = 0 or (y – x)(xy + 1) = 0
1
either y = x or y=−
x
First putting y = x in (1), 2x3 – 18x = 0 or 2x(x2 – 9) = 0
implying x = 0, 3, –3. …(3)
Hence the stationary points are (0, 0), (3, 3), (–3, –3).
1
Again putting y=− in (1),
x
2
2x − − 10x − 8 − = 0
1 1 10
or − 10x or x2 = 1
x x x
implying x = ±1 …(4)
Hence stationary points are (1, –1), (–1, 1).
Thus total number of stationary points are (0, 0), (3, 3), (–3, –3), (1, –1), (–1, 1).
At (0, 0): (rt − s2 ) = fxx fyy − fxy
2
= (2y2 − 10)(2x2 − 10) − (4xy − 8)2 = 36 > 0
and r(0, 0) = (2y2 – 10) = –10 < 0
whence f(x, y) has maximum at (0, 0).
At (3, 3): (rt – s2) = (2y2 – 10)(2x2 – 10) – (4xy – 8)2 = 64 – (28)2 < 0
and r = 2y2 – 10 = 8 > 0
whence it is a point of minimum.
At (–3, –3): (rt – s2) = (2y2 – 10)(2x2 – 10) – (4xy – 8)2 = (18 – 10)(18 – 10) – (36 – 8)2 < 0
whence at (–3, –3), f(x, y) is neither maximum nor minimum.
At (1, –1): (rt – s2) = (2y2 – 10)(2x2 – 10) – (4xy – 8)2 < 0 (negative)
whence again at (–1,–1), function has neither maximum nor minimum.
At (–1, 1): (rt – s2) < 0. Therefore the function has neither maximum nor minimum at (–1, 1).
332 Engineering Mathematics through Applications
4.11 LAGRANGES METHOD OF UNDETERMINED MULIPLIERS: CONSTRAINED
MAXIMA-MINIMA
In many engineering and science problem it is desired to find extrema of a function of
several variables that are not all independent but are connected to one another by certain
conditions. In such cases, generally the conventional method becomes either very complex
or impracticable, then we employ an alternate method which is very easy in its approach,
called Lagrange’s method of undetermined multipliers. Another name of it viz. constrained
maxima-minima is very obvious as here the variables involved in the function of which
extreme values are to be obtained are linked to each other by certain relations which are to
be taken care in the process of finding extreme values.
Illustration: Say, if we want to find the Maximum and Minimum values of
u = f(x, y, z) …(1)
Consider function of three variables x, y, z which are connected by an implicit relation,
φ(x, y, z) = 0 …(2)
For function u to possess stationary values, it is necessary that
∂u ∂u ∂u
= 0, = 0, =0 …(3)
∂x ∂y ∂z
∂u ∂u ∂u
∴ dx + dy + dz = du(x, y, z) = 0 …(4)
∂x ∂y ∂z
Also from (2),
∂φ ∂φ ∂φ
dx + dy + dz = dφ = 0 …(5)
∂x ∂y ∂z
We see that (3) + λ(4) results in
∂u ∂φ ∂u ∂φ ∂u ∂φ
+ λ dx + + λ dy + + λ dz = 0 …(6)
∂x ∂x ∂y ∂y ∂z ∂z
but this will hold true only if
∂u ∂φ
+λ = 0, (i)
∂x ∂x
∂u ∂φ
+λ = 0, (ii) …(7)
∂y ∂y
∂u ∂φ
+λ = 0, (iii)
∂z ∂z
whence these three equation of (7) taken together with (2) will determine those x, y, z and λ
for which u is a stationary.
Observation: Though this method is very simple in its approach, but it fails to determine the nature of the
stationary points whether they are points of maximum or minimum or saddle points.
Partial Derivatives and their Applications 333
Working Rule:
1. Write an auxiliary function F(x, y, z) = u(x, y, z) + λφ(x, y, z).
∂F ∂u ∂φ
=0= + λ ,
∂x ∂x ∂x
∂F ∂u ∂φ
2. Meet necessary conditions viz =0= + λ ,
. ∂y ∂y ∂y
∂F ∂u ∂φ
=0= +λ
∂z ∂z ∂z
3. Solve the above equations together with φ(x, y, z) = 0.
Example 74: Find the dimensions of a rectangle box, open at the top, of maximum capacity
whose surface is 432 sq.cm. [MDU, 2004]
Solution: Let x, y, z be the three dimension of the box so that Volume, V(x, y, z) = xyz
Now our object is to find such values of x, y, z for which V(x, y, z) is maximum with the given
condition that
Surface area, S(x, y, z) = xy + 2yz + 2zx = 432 sq. cm …(1)
Define Lagrange’s Function,
F(x, y, z) = V(x, y, z) + λ S(x, y, z) = xyz + λ(xy + 2yz + 2zx – 432) …(2)
For stationary values, dF = 0, i.e.
∂F
= 0 i.e. yz + λ(y + 2z) = 0 (i)
∂x
∂F
= 0 i.e. xz + λ(x + 2z) = 0 (ii)
∂y …(3)
∂F
= 0 i.e. xy + λ(2x + 2y) = 0 (iii)
∂z
In (3), multiply (i) by x, (ii) by y, (iii) by z and add all, we get
λ(2xy + 4xz + 4yz) + 3xyz = 0
xyz
2λ(432) + 3xyz = 0 or λ=− …(4)
288
V
Substituting of λ = − in (3),
288
(y + 2z) = 0 (iv)
x
1−
288
y
1− (x + 2z) = 0 (v)
288 …(5)
z
1− (2x + 4z) = 0 (vi)
288
72
From (iv) and (v), we get x = y, which on putting in (vi) gives, z = .
y
334 Engineering Mathematics through Applications
72
Now putting x = y, z = , into (iv),
y
y 72
1− y + 2 ⋅ = 0 or 288 – (y2 + 144) = 0 or y2 = 144 means y = 12 cm
288 y
72 72
y = 12 means so is x = 12 and = z=
= 6 cm
y 12
Hence x = 12 cm, y = 12 cm, z = 6 cm gives the maximum volume.
Example 75: If u = a3x2 + b3y2 + c3z2, where x–1 + y–1 + z–1 = 1, show that the stationary value
of u is given by x = ∑ a , y = ∑ a , z = ∑ a . [UP Tech, 2004; KUK, 2007; PTU, 2008]
a b c
Solution: Consider
F(x, y, z) = u(x, y, z) + λ φ(x, y, x) = (a3x2 + b3y2 + c3z2) + λ(x–1 + y–1 + z–1) …(1)
Then for stationary values,
∂F ∂u ∂φ λ
= +λ =0 2a3 x + = 0 or 2a3 x3 = λ, (i)
∂x ∂x ∂x x2
∂F ∂u ∂φ λ
= +λ =0 2b3 y + 2 = 0 or 2b3 x3 = λ , (ii)
∂y ∂y ∂y y …(2)
∂F ∂u ∂φ λ
= +λ =0 2c3 z + 2 = 0 or 2c3 z3 = λ, (iii)
∂z ∂z ∂z z
From (i), (ii), (iii),
1
λ
ax = by = cz =
3
2a3x3 = 2b3y3 = 2c3z3 = λ or = k (say)
2
(i)
k
x=
a
k
y= (ii)
or b …(3)
k
z= (iii)
c
1 1 1 a b c
so that + + = 1 gives + + = 1 implying (a + b + c) = k …(4)
x y z k k k
k a + b + c ∑a k ∑a k ∑a
Therefore, from (3), x = = = , y= = , z= = ,
a a a b b c c
Example 76: Find the volume of the greatest rectangular parallelopiped that can be
x2 y2 z2
inscribed inside the ellipsoid + + =1
a2 b2 c2
Partial Derivatives and their Applications 335
Solution: Let edges of the parallelopiped be 2x, 2y, 2z parallel to the coordinate axes so that
volume,
V = 8xyz …(1)
Our object is to maximise V(x, y, z) subject to the condition,
x2 y2 z2 x2 y2 z2
+ + = 1 or φ(x, y , z) = + + −1= 0 …(2)
a2 b2 c2 a2 b2 c2
Define function,
x2 y2 z2
F(x, y, z) = V + λφ = 8xyz + λ 2 + 2 + 2 − 1 …(3)
a b c
So that for stationary values,
∂F 2x
= 8yz + λ 2 = 0 …(i)
∂x a
∂F 2y
= 8xz + λ 2 = 0 …(ii)
∂y b …(4)
∂F
…(iii)
2z
= 8xy + λ 2 = 0
∂z c
Equating the values of λ from (i) and (ii); (i) and (iii), we get
x2 y2 x2 z2
= and =
a2 b2 a2 c2
x2 y2 z2 1 using x2 = y = z2
2
= = = a2 b2 c2
implying thereby ; in (2) …(5)
a2 b2 c2 3
a b c
or x= , y= , z=
3 3 3
When x = 0, the parallelopiped mearly becomes a rectangular sheet and in that case the
volume, V = 0.
a b c 8 abc
Hence V is maximum when x = ,y= ,z= and VMaxi = 8xyz = .
3 3 3 3 3
Example 77: Find the maximum value of the function cosA cosB cosC.
Solution: If A, B, C are the angles of a triangle ABC, then
A + B + C = 180° (an implicit relation) …(1)
For f (A, B, C) = cosA cos B cosC, …(2)
define Lagrange’s function,
F(A, B, C) = f + λφ = cosA cosB cosC + λ(A + B + C – π) …(3)
336 Engineering Mathematics through Applications
Applying maximization conditions,
∂F
= − sin A cos B cos C + λ = 0 …(i)
∂A
∂F
= − cos A sin B cos C + λ = 0 …(ii)
∂B …(4)
∂F
= − cos A cos B sin C + λ = 0 …(iii)
∂C
From (i) and (ii),
sinA cosB cosC = cosA sin B cosC or cosC (sinA cosB – sinB cosA) = 0
Implying either sin (A – B) = 0 i.e. A = B
π …(5)
or cosC = 0 i.e. C= .
2
Likewise from (ii) and (iii),
cosA sinB cosC = cosA cosB sinC or cosA(sinB cosC – cosBsinC) = 0
Implying either sin(B – C) = 0 i.e. B = C
π …(6)
or cosA = 0 i.e. .A=
2
Taking (iv) and (v) together, we conclude that A = B = C = 60° as two of the angles in a
π
triangle cannot be 2 each.
π
Here f = cosA cosB cosC is maximum at A = B = C = and fMaxi = (cos 60°)3 = 1 .
3 8
Alternate: By Conventional Method, f(A, B, C) = cosA cosB cosC
On using the condition, A + B + C = π or C = [π – (A + B)],
reduces to a function of two variables
i.e. f(A, B, C) = cosA cosB cos[π – (A + B)] = –cosA cosB cos(A + B) …(7)
and then for stationary values,
∂f
= sin A ⋅ cos B cos(A + B) + cos A cos B sin(A + B) = 0, …(8)
∂A
∂f
= cos A sin B cos(A + B) + cos A cos B sin(A + B) = 0 …(9)
∂B
From (8),
cosB[sinA cos(A + B) + cosA sin(A + B)] = 0 or cosB sin(2A + B) = 0
π
implying B= , 2A + B = π …(10)
2
From (9),
cosA sinB cos(A + B) + cosA cosBsin(A + B) = 0 or cosA sin(A + 2B) = 0
π
implying, A= , A + 2B = π …(11)
2
Partial Derivatives and their Applications 337
π π
Together, (11) and (12) gives A = B = as A = is not possible and so is the third angle C.
3 2
Further find r = fAA, s = fAB , t = fBB and observe signs of (rt – s2) and r to ascertain,
π
A=B=C= .
3
Example 78: Find the maximum and minimum distances from the origin to the curve
5x2 + 6xy + 5y2 – 8 = 0.
Solution: Here we need to determine the extreme values of the function f(x, y) = x2 + y2
(or d = x2 + y2 ) subject to the condition φ(x, y) = 5x 2 + 5y2 + 6xy – 8 = 0
Thus,
F(x, y) = f(x, y) + λ φ(x, y) = (x2 + y2) + λ(5x2 + 5y2 + 6xy – 8) …(1)
∂F
Now = 2x + λ(10x + 6y) = 0 …(2)
∂x
∂F
and = 2y + λ(6x + 10y) = 0 …(3)
∂y
In order to solve (2) and (3) for x and y, multiply (2) by y and (3) by x and then substract
the later from the former,
6λ(y2 – x2) = 0 i.e. y = ± x …(4)
On substituting y = ±x into the relation, 5x + 6xy + 5y – 8 = 0, we get,
2 2
2x2 = 1 …(i) …(5)
and x2 = 2 …(ii)
Thus, the distance, f(x, y) = (x2 + y2) from the origin becomes
1 1 1
d2 = x2 + y2 = + = 1, using x2 = = y2
2 2 2
and d2 = x2 + y2 = 2 + 2 = 4, (using x2 = 2 = y2)
Obviously, the first value is a minimum whereas second value is a maximum. Clearly, the
curve is an ellipse with semi-axes 2 and 1, whose major axis makes an angle of 45° with the X-
axis.
Example 79: Prove that the rectangular solid of maximum volume that can be inscribed in
a given sphere is a cube.
Solution: Let x, y, z be the edges of the rectangular solid so that
V = xyz …(1)
Since the rectangular solid is inscribed in sphere, then the diagonal of the rectangular
solid is the diameter of the sphere.
∴ x2 + y2 + z2 = d or S(x, y, z) ≡ (x2 + y2 + z2 – d2) …(2)
338 Engineering Mathematics through Applications
Now the condition for the volume of the rectangular solid to be maximum is that
∂V ∂S ∂V ∂S ∂V ∂S
+λ = 0, +λ = 0, +λ =0 …(3)
∂x ∂x ∂y ∂y ∂z ∂z
yz + 2λx = 0 …(i)
i.e. zx + 2λy = 0 …(ii) …(4)
xy + 2λz = 0 …(iii)
Multiply (i) by x, (ii) by y, (iii) by z and equate them
From (i) and (ii), we get
x=y
From (i) and (iii), we get x = z }
i.e. x = y = z …(5)
Hence the given rectangular solid would be of maximum volume if it is a cube.
Alternately (Conventional Method)
From S ≡ x2 + y2 + z2 – d2, let us consider z as a function of x and y.
∂S ∂z ∂S ∂z
= 2x + 2z …(i) = 2y + 2z
∂x ∂x ∂y ∂y …(ii)
(³ y is independent of x) (³ x is independent of y)
∂S ∂S ∂z x ∂z y
Here =0= implies =− and =− …(6)
∂x ∂y ∂x z ∂y z
∂z x2 y x2
Now Vx (x, y, z) = 1 ⋅ yz + xy = yz − = yz − = 0 …(7)
∂x z z
∂z xy2 y2
and Vy(x, y, z) = 1 ⋅ xz + xy = xz − = x z − = 0 …(8)
∂y z z
(7) and (8) collectively implies x = y = z.
Further to check whether these points are the points of maxima, we find vxx, vxy, vyy.
∂z 2 x
∂z
2zx − x2
∂ x x 2zx − x − z − xy(x2 + 3z2 )
From (7), r = y − = y − z − = = −4x
∂x z2 z2 z3
(at x = y = z) …(9)
x2 ∂z x2 ∂z x2 y x2 y
From (7), s = 1 z − + y + 2 = z − + y − + 2 − = −2x ,
z ∂y z ∂y z z z z
(at x = y = z) …(10)
−xy(y2 + 3z2 )
From (8), t= = −4x, (at x = y = z) …(11)
z3
rt – s2 = (–4x)(–4x) – (–2x)2 = 12x2 > 0 and r = –2x < 0 for all x.
Hence x = y = z is the point of maximum, V(x, y, z) = xyz = x3, i.e. a cube.
Partial Derivatives and their Applications 339
Example 80: Find the stationary values of (x2 + y2 + z2) subject to ax2 + by2 + cz2 = 1 and
lx + my + nz = 0
Solution: Let f(x, y, z) = (x2 + y2 + z2) …(1)
and φ(x, y, z) = (ax2 + by2 + cz2 – 1) = 0 …(2)
ψ(x, y, z) = (lx + my + nz) = 0 …(3)
Then by Lagranges Multiplier,
F(x, y, z) = (x2 + y2 + z2) + λ(ax2 + by2 + cz2) + 2µ(lx + my + nz)
which implies
dF = (2x + λ2ax + 2µl)dx + (2y + λ2by + 2µm)dy + (2z + λ2cz + 2µn)dz
Then for F(x, y, z) to possess stationary values,
∂F
= x + λax + µl = 0 …(i)
∂x
∂F
= y + λby + µm = 0 …(ii) …(4)
∂y
∂F
= z + λcz + µn = 0 …(iii)
∂z
In the above results, first multiply (i) by x, (ii) by y, (iii) by z and then add the three,
(x2 + y2 + z2) + λ(ax2 + by2 + cz2) + µ(lx + my + nz) = 0
On using (1), (2), (3), the above expression becomes
f + λ · 1 + µ · 0 = 0 i.e. λ = –f …(5)
µl
Thus from 4(i), x + (–f)ax + µl = 0 or x= ,
(af − 1)
µm
Similarly from, 4(ii) and 4(iii), y= ,
(bf − 1) …(6)
µn
z=
(cf − 1)
Substituting the values of x, y, z again in the relation (3), we get the condition
µl µm µn
l + m + n =0
af − 1 bf − 1 cf − 1
l2 m2 n2
As µ ≠ 0, + + = 0, …(7)
af − 1 bf − 1 cf − 1
This is a quadratic in f from which we can obtain its extreme values.
Further, the surface ax2 + by2 + cz2 = 1 is an ellipsoid or a hyperboloid and lx + my + nz = 0
is a plane. Therefore point (x, y, z) satisfying both these equations lies on the conic of their
intersection and the expression (x2 + y2 + z2) gives us the square of the distance of (x, y, z)
from the origin.
The maximum and minimum values of this distance are the major and minor axes of the
same conic of intersection. Hence the equation (7) gives the squares the lengths of the semi-
axes of the conic of intersection.
340 Engineering Mathematics through Applications
Example 81: A tent on a square base of sides x, has its sides T
vertical of height y and the top is a regular pyramid of height h
h. Find x and y in terms of h, if the canvas required for its G
H
construction is to be minimised for the tent to have a given
capacity. M
E F
Solution: Let the tent (with cuboid ABCD, EFGH and
Pyramid T EFGH atop) be made on a square base ABCD with
side x (as in Fig. 4.6). y
Here V(x, y, h) = Volume of the box (cuboid) + Volume
D
due to elevated Portion (pyramid) C
x h = x2 y +
1 2 h x
= x2 y + …(1)
3 3 A x B
and S(x, y, h) = Area of the four faces of the cuboid + Area Fig. 4.6
of the four faces of the pyramid.
2
1 x 2 x
= 4xy + 4 ⋅ base × length TM = 4xy + 4 ⋅ h + = 4xy + x x2 + 4h2 …(2)
2 2 2
Define Lagrange’s Function,
h
F(x , y, h) = S(x, y , h) + λV(x , y, z) = 4xy + x x2 + 4h2 + λ x2 y + …(3)
3
For surface area to be minimum (subject to given capacity),
∂F h
⋅ 2x + λ 2x y +
1 1
= 4y + x2 + 4h2 + x = 0 …( i)
∂x 2 x + 4h
2 2
3
∂F
= 4x + λ ⋅ x2 = 0 …(ii) …(4)
∂y
∂F 4hx x2
= +λ =0 …(iii)
∂h x + 4h
2 2 3
4 4hx 4x
4(ii) gives λ = − which on putting in 4(iii) gives − or x2 = 5h2
x x + 4h
2 2 3
4 4
On putting into 4(i), x = 5 h and λ=− =− , we get
x 5h
1 1 4 h
4y + 5h + 4h + 5 h 2 ⋅ 2 5 h − 2 5h y +
2 2
5h + 4h
2 2
5 h 3
4y + 3h + 5 h − 8 y + h = 0 or –4y + 2h = 0, i.e. y =
h
3 3 2
Partial Derivatives and their Applications 341
Alternately: By Conventional Method;
x2 h
V−
3 = V − xh
V = x2 y + = x(xy) +
2
h xh
or xy = …(5)
3 3 x x 3
1
S = 4xy + x x2 + 4h2 = 4 − + x (x2 + 4h2 )2
V xh
x …(6)
3
Thus by elimination of y, surface area S has been made a function of two variables x and
h only.
Now for minimum surface area, Sx = 0 = Sh.
Therefore, Sx (x, h) = −4V ⋅ 2 − h + x2 + 4h2 + x2
1 4
2
=0 …(7)
x 3 x + 4h
2
4 4xh
and Sh (x, h) = − x + =0 …(8)
3 x + 4h2
2
4 4xh
From(8) − x+ =0 or (x2 + 4h2) = 9h2 or x = 5h
3 x2 + 4h2
Putting x = 5 h into (7),
4 2 h 4 5h2 2 h
− 2 x y + − h + 3h + = 0, as V = x y +
x 3 3 3h 3
4 4 5 h
or −4y − h − h + 3h + h = 0 or − 4y − 2h = 0 ⇒ y = .
3 3 3 2
ASSIGNMENT 9
1. Find the maximum and minimum values of
a3 a3
(i) xy + + (ii) 2(x2 – y2) – x4 + y4 [NIT Kurukshetra, 2005]
x y
2. Find the minimum value of x2 + y2 + z2, given that ax + by + cz = p
3. Find the dimensions of a rectangle box without top with a given capacity so that the
material used is minimum.
4. Determine the maxima of the function given by u = (x + 1)(y + 1)(z + 1)
subject to the condition xaybzc = k.
[Hint: Take log of both functions]
5. Divide 24 into three parts such that the continued product of the first, square of the
second and the cube of the third may be maximum.
6. Given x + y + z = a, find the maximum value of xm yn zp. [KUK, NIT Kurukshetra, 2004]
7. Find the point on the surface z2 = xy + 1 nearest to the origin.
342 Engineering Mathematics through Applications
8. The temperature T at any point (x, y, z) in space is T = 40xyz2. Find the highest temperature
on the surface of the unit sphere x2 + y2 + z2 = 1.
9. Find a point within a triangle such that the sum of the squares of its distances from the
three vertices is minimum.
or
Find the point upon the plane ax + by + cz = 0 at which the function φ = x2 + y2 + z2 has a
minimum value and find this minimum φ.
x2 y2 z2
10. Show that stationary values of u = + + , where lx + my + nz = 0 and
a4 b4 c4
x2 y2 z2 l2 a4 m2 b4 n2 c4
+ + = 1 are the roots of the equation + + = 0.
a2 b2 c2 1 − a u 1 − b u 1 − c2u
2 2
11. Show that if the parameter of a triangle is constant, its area is a maximum when it is
equilateral. [NIT kurukshetra, 2008]
[Hint: f = ∆ = s(s – a)(s – b)(s – c), where a + b + c = 2s (s const.)]
2
12. Find the triangle of maximum area inscribed in a circle.
r2
Hint : Take Area S( A, B, C) = (sin 2A + sin 2B + sin 2C) ⋅ 2 , where A + B + C = π
4.12 DIFFERENTIATION UNDER INTEGRAL SIGN
If f(x, α) a function of two variables with α as parameter, be integrated with respect to x
∫
b
between the limits a and b, then f (x, α) dx will be a function of α: F(α), say.
a
d b
∫
d
To find F(α) = f (x, α) dx , when it exists, it is not always possible to find the
d α α
d a
integral first and then to find its derivative. Such problems are handled by the following
rules:
Leibnitz’s Rule* 1: (When Limits of integration are constant)
∂
Statement: If f(x, α) and f (x, α) be continous function of x and a, then
∂α
d
∫ ∂
b
∫ ∂α f (x, α) dx, where a and b are constants independent of α.
b
f (x, α)dx =
dα a a
∫
b
Poof: Let f (x, α)dx, then for small change α to α + δα,
a
b b b
F(α + δα) − F(α) = ∫ f (x, α + δα) dx –∫ f (x, α) dx =∫ f (x, α + δα) − f (x, α) dx
a a a
On using Mean value theorem of differential calculus as staled below
∂
f (x, α + δα) − f (x, α) = δα f (x, α + θh), where 0 < θ < 1.
∂α
*Named after the German mathematician Gottfried Wilhelm Leibnitz (1646–1716).
Partial Derivatives and their Applications 343
∂
∫ ∂α f (x, α + θh) dx
b
We get, F(α + δα) − F(α) = δα ⋅
a
F(α + δα) − F(α) ∂
∫ ∂α f (x, α + θ ⋅ 0)dx
b
Taking limit as δα → 0, lt =
δα→ 0 δα a
∂
∫ ∂α f (x, α) dx
b
d
or F(α) =
dα a
Leibnitz’s Rule 2: (When Limits of integration are function of parameter)
∂
If f(x, α) and ∂α f (x, α) are continuous function of x and α, then
d ψ(α)
ψ( α )
∂ f (x, ) dx dψ f ( ), dφ
∫ ∫ ( ψ α α) − f ( φ(α), α )
d
F(α) = f ( x, α) dx = α +
dα dα φ(α) φ(α) ∂α dα dα
provided φ(α) and ψ(α) possesses continuous first order derivatives with respect to α.
∫
x
1 1 x
Example 82: Differentiating dx = tan−1 under the integral sign, find the value
0 x2 + a2 a a
∫
x
1
of dx
0 (x2 + a2 )2
= tan−1 ∫
x
dx 1 x
Solution: Differentiating both sides of
0 (x + a )
2
a a 2
with respect to ’a’ (the parameter) by Leibnitz’s Rule 1,
∂ 1 ∂ 1 −1 x
∫
x
2 2
dx = ⋅ tan
0 ∂a x + a ∂a a a
1 ∂ x ∂ 1
∫
x
1 x
− ⋅ 2a dx = tan−1 + ⋅ tan−1
implying 0 (x2 + a2 )2 a ∂a a ∂a a a
−x
∫
x
1 1 x
i.e. − 2a ⋅ dx = − 2 tan−1
0 (x + a )
2 2 2
a(a + x ) a
2 2
a
Dividing throughout by –2a,
∫
x
1 1 x x
dx = 3 tan −1 + 2 2
0 (x2 + a2 )2 2a a 2a (a + x2 )
∞
e− ax sinmx
∫
m
Example 83: Prove that if a > 0, dx = tan−1 .
0 x a
∞ − ax
∫
e sin mx
Solution: Let F(m) = dx …(1)
0 x
344 Engineering Mathematics through Applications
By Leibnitz’s Rule 1,
∞ ∞
∂ e−ax sin mx ∂ e−ax
∫ ∫
d
F(m) = dx = (sin mx) ⋅ dx
dm 0 ∂m x 0 ∂m x
∞
e−ax ∞
=
∫ 0
x(cos mx)
x
dx = ∫
0
e− ax cos mx dx
∞
dF e−ax a
= ( − a cos mx + m sin mx) = 2 …(2)
dm (−a) + m
2 2 +
0 a m
2
∞
∫
eax
using eax cos bx dx = (a cos bx + b sin bx)
0 (a + b )
2 2
Integrating both sides with respect to x,
∫a dx = a tan −1 + C = tan −1 + C, a > 0
1 1 m m
F(m) = a …(3)
2
+ m2 a a a
When m = 0; from (3), F(0) = tan−1 0 + C = C implying C = 0
from (1), F(0) = 0
∞ − ax
∫
e sin mx m
∴ dx = tan−1 .
0 x a
∞ −x
∫
e
Example 84: Prove that (1 − e−ax )dx = log(1 + a), (a > – 1)
0 x
∞ −x
∫
e
Solution: Let I = (1 − e−ax ) dx …(1)
0 x
∞ ∞ −x
∂ e−x ∞
∫ ∫ ∫
dI e
so that = (1 − e−ax ) dx = ⋅ e−ax (−x) dx = e−x e−ax dx
da 0 ∂a x 0 x 0
∞
∞ e−(1+ a)x
∫
dI 1
= e−(1+ a)x dx = = (1 + ) , a > − 1
da 0 −(1 + a) 0 a
On integration,
I = log (1 + a) + C …(2)
∞ −x
∫
e
Now when a = 0 then from (1), I = (1 − 1) dx = 0 thereby implying C = 0
0 x
from (2), I = log(1 + 0) + C = C
∴ I = log (1 + a), a > – 1.
π/2 log (1 + y sin2 x)
Example 85: Show that ∫ 0 sin2 x
dx = π ( 1+ y −1 ) [NIT Kurukshetra, 2002]
Partial Derivatives and their Applications 345
π
log (1 + y sin2 x)
∫
2
Solution: Let I= dx …(1)
0 sin2 x
By Leibnitz’s Rule 1,
π π
2 ∂ log (1 + y sin2 x) ∂
∫ ∫
dI 2 1 1
= dx = ⋅ (y sin2 x) dx
da 0
∂y sin2 x 0 sin x (1 + y sin x ) ∂y
2 2
π π
∫ ∫
2 1 2 1
= dx = dx
0 (1 + y sin2 x) 0 cos x + sin2 x + y sin2 x
2
π π
∫ ∫
2 1 2 sec2 x
= dx = dx …(2)
0 cos x + (1 + y)sin2 x
2 0 1 + (1 + y)tan2 x
Putting tan x = t so that sec2x dx = dt, and for x = 0; t = 0, x = π 2 , t = ∞
∞ ∞ ∞
∫ ∫
dI 1 1 1 1 1 t
= dt = dt = tan−1
∴ da 0 1 + (1 + y) t2
(1 + y) 0 1
2
(1 + y) 1 1
1+ y
1+ y + t
2
1 + y
0
π
1
dI 1 −
= tan−1 ∞ − tan−1 0 = (1 + y) 2 …(3)
dy 1+ y 2
Integrating (3) with respect to y,
1
π (1 + y) 2 1
I= = π(1 + y) 2 +C
2 1 …(4)
2
When y = 0 from (1), I(0) = 0, thereby implying C = –π
from (4) I(0) = π + C
1
Hence I = π(1 + y) 2 − π = π 1 + y − 1 .
π
Example 86: Evaluate ∫ 0
log (1 + a cos x) dx using the method of differentiation under
integral sign.
OR
π
∫ log (1 + a cos x) dx = π log + 1 − a2 .
1 1
If |a| < |, prove that
2 2 2
π
Solution: Let I(a) = ∫ log(1 + a cos x) dx
0
π π
∂ log(1 + cos ) =
∫ ∫
dI cos x
so that = ∂a a x dx dx
da 0 0 1 + a cos x
346 Engineering Mathematics through Applications
π
1 + a cos x − 1 π π
∫ ∫ ∫
1 1 1 1
= dx = dx − dx
a 0 1 + a cos x a 0 a 0 1 + a cos x
π π β + α cos x
∫ ∫
1 1 1
dx, use dx = cos−1 taking α = 1, β = a.
For 0 1 + a cos x 0 α + β cos x α −β
2 2 α + β cos x
π
dI π 1 1 −1 a + 1cos x π 1 1
∴ = − cos 1 + a cos x = a − a 1 − a2 cos (− 1) − cos 1
−1 −1
da a a 1 − a2 0
π 1 1 π 1
= − (π − 0) = 1 − .
a a 1− a2 a 1 − a2
Integrating both sides with respect to a,
π
∫ a da − π∫ a
1 1
I=π da + C
1 − a2
∫ sin t 1 − sin t cos t dt + C (Put a = sint, da = cost dt)
1
= π log a − π 2
= π log a − π∫
1
dt + C = π log a − π∫ cosec t dt + C
sin t
= π loga + π log (cosec t + cot t)
1 + cos t 1 + 1 − sin2 t
= π log a + π log + C = π log a + π log +C
sin t sin t
I = π log a + π log
1 + 1 − a2
a
+ C = π log 1 + 1 − a2 + C ( )
When a = 0, from above I(0) = π log 2 + C
implying C = – π log2.
from given integral, I(0) = 0
1 + 1 − a2
∴ (
I = π log 1 + 1 − a2 − π log 2 = π log
) 2
∞
π − α4 ∞
π α
2 2
∫ ∫
−
e− x cosαx dx = x ex sinαx dx =
αe 4
2 2
Example 87: Prove that e and
0 2 0 4
[NIT Kurukshetra, 2005; KUK, 2009]
∞
∫ e− x cos αx dx =I(α)
2
Solution: Let …(1)
0
∞
∂ −x2
∫
dI
so that = (e cos αx) dx
dα 0 ∂x
∞
=∫ e−x (− sin αx) x dx
2
0
Partial Derivatives and their Applications 347
( )
∞
∫
dI 1 (sin αx)(−2x e−x ) dx, write − 2xe−x = d e−x
=
2 2 2
dα 2 0
I II
On integrating right hand expression by parts, we get
{ } −∫
1 ∞
dI ∞
= sin αx (e−x ) α cos αx ( e−x ) dx
2 2
dα 2 0 0
1 ∞
∫
dI
= 0 − α e−x cos αx dx
2
⇒
dα 2 0
dI α
⇒ =− I …(2)
dα 2
dI
⇒ dα = − α
I 2
α2 α 2
∫
− f'(x)
log I = −+ log C or I = C e 4 using dx = log f (x) …(3)
4 f (x)
But when α = 0, then from (1),
∞
π (Gama function)
I= ∫ e−x dx =
2
…(4)
0 2
π
From (3), when α = 0, we get C = …(5)
2
π2
π −4
and I=
e …(6)
2
Thus the given integral
∞ α2
π −4
∫ e−x cos α x dx =
2
e …(7)
0 2
On differentiating both sides with respect to α, we get
π − α4 α
2
∞
∫ e−x (− sin αx) x dx = −
2
e
0 2 2
∞ α2
π
∫
−
x e −x sin α x dx = αe
2
i.e. 4
0 4
Hence the result.
∞
tan−1ax π
Example 88: Show that ∫ 0 x(1 + x )
2
dx = log(1 + a), a ≥ 0
2
[KUK, 2000, 2005; NIT Kurukshetra, 2004]
348 Engineering Mathematics through Applications
∞
tan−1 ax
Solution: Let I= ∫0 x(1 + x2 )
dx …(1)
Differentiating both sides with respect to a, we have
∂ ∞
tan−1 ax
∫
dI
= dx
da ∂ a 0 x(1 + x2 )
∞
∂ tan −1
∫
1
= ax dx
0 x(1 + x2 ) ∂a
∫
1 x
= ⋅ dx
0 x(1 + x ) 1 + a2 x2
2
∫
dI 1
= dx …(2)
da 0 (1 + a2 x2 )(1 + x2 )
Dealing the integrand by making partial fractions on taking x2 = t, we get
1 A B
= +
(1 + a2 t)(1 + t) (1 + t) (1 + a2 t)
⇒ 1 = A(1 + a2t) + B(1 + t) …(3)
When 1 + a2t = 0, we get B 1 − 2 = 1 i.e. B = − a 2
1 2
a 1− a
1
When 1 + t = 0, we get 1 = A(1 – a2), i.e. A =
(1 − a2 )
Therefore on substituting values of (A) and (B) in (3), we get
1 1 1 a2
= − , 2
(1 + a t) (1 + t) (1 − a )(1 + t) (1 + a t)(1 − a2 ) t = x
2 2 2 …(4)
Using (4), integrand (2) becomes
∞
1 1 a2
∫
dI
= −
(1 − a2 ) (1 + x2 ) 1 + a2 x2
dx
da 0
∞ ∞
1
∫ ∫
dI 1 1
= dx − dx
da (1 − a2 ) 0 (1 + x2 ) 0
2
1 + 2
or x
a
( tan−1 x ) − ( a tan−1 ax )
1 ∞ ∞
= 2
(1 − a ) 0 0
1 π π (1 − a) π π
− 0 − a − 0 =
dI
= 2
=
da (1 − a ) 2 2 (1 − a2 ) 2 2(1 + a)
Partial Derivatives and their Applications 349
On integrating both sides with respect to a,
π
or I=log(1 + a) + C …(5)
2
Now on using conditions, when a = 0, I = 0 (from (1))
π
⇒ 0= log(1 + 0) + C , i.e. C = 0
2
With C = 0, (5) becomes
π
I= log(1 + a)
2
∞ − x2 + a2
2
Example 89: Evaluate
∫ 0
e x
dx by differentiating under the integral sign.
[NIT Kurukshetra, 2003]
∞ − x2 + a2
Solution: Let F(a) =
∫ 0
e x2 dx …(1)
By beibnitz Rule 1,
a2 ∞ − x2 + a
2
∞
∂ − x2 + x2 − 2a dx
∫ ∫
d x2
F(a) = e dx = e 2
da 0 ∂a 0 x
a a
Put x = so that dx = − 2 dt and limits for x → 0, t → ∞; x → ∞, t → 0.
t t
0 − a + t2 0 − t2 + a
2 2
− 2at2 − a
∫e ∫e
d t2 t2
∴ F(a) = 2 ⋅ 2 ⋅ dt = 2 dt
da ∞ a t ∞
∞ − x2 + a
2
∫
dF(a) x2 dF(a)
= −2 e dx = −2F(a) or = −2 da …(2)
da 0 F(a)
On integration,
log F(a) = – 2a + logC or logF(a) = log e– 2a + logC
implying F(a) = Ce– 2a …(3)
When a = 0, from (3), F(0) = C
∞
π giving C = π
from (1), F(0) = ∫0
e − x2
dx =
2 2
…(4)
∞ − x2 + a
2
π −2 a
Using (4), F(a) = ∫ 0
e x2 dx = Ce−2 a =
2
e .
α log (1 + αx) log (1 + x) π
∫ ∫
1
Example 90: Evaluate dx and hence show that dx = loge 2.
0 (1 + x2 ) 0 (1 + x2 ) 8
350 Engineering Mathematics through Applications
α log(1 + αx)
Solution: Let F(α) = ∫ 0 (1 + x2 )
dx …(1)
By Leibnitz’s Rule 2,
α
∂ log(1 + α2 ) log(1 + α2 )
∫
dF d
= dx + (α)
dα ∂x 1 + α (1 + α2 )
2
0 dx
α log(1 + α2 )
∫
x
= dx + …(2)
0 (1 + α x)(1 + x2 ) 1 + α2
Resolving the integrand into partial fractions:
x A Bx + C
= +
(1 + αx)(1 + x ) 1 + αx 1 + x2
2
implying x = A(1 + x2) + (Bx + C)(1 + αx) or x = (A + Bα)x2 + (B + Cα)x + (A + C)
−1 −1 −α
, we get α = A 1 + 2
1
Putting 1 + αx = 0 i.e. x= or A=
α α 1 + α2
−A 1
Comparing coefficients of x2, 0 = A + αB or B= =
α 1 + α2
α
Comparing constant terms, 0 = A + C or C = −A =
1 + α2
α
1 α −α α α
∫ ∫ ∫ ∫
x 1 2x 1
dx = dx + dx + α
(1 + α2 )
∴ dx
0 (1 + αx)(1 + x )
2 0 1 + αx 2 0 (1 + x2 ) 0 1+ x2
α
1 −α log(1 + αx) 1
= + log(1 + x2 ) + α tan− 1 x
(1 + α2 ) α 2 0
α
1
∫ − log(1 + α2 ) + log(1 + α2 ) + α tan− 1 α
x 1
dx =
0 (1 + αx)(1 + x2 ) (1 + α2 ) 2 …(3)
With the above integral value, (2) on cancellation of some terms reduces to
dF 1 log(1 + α2 ) α
= + tan −1 α …(4)
dα 2 (1 + α )
2
(1 + α )
2
Integrating (4) with respect to a,
α
∫ ⋅ log(1 + α2 ) dα + ∫
1 1
F(α) = (tan−1 α) dα + C (Integration by parts)
2 (1 + α2 ) 1 + α2
1 α α
= log(1 + α2 )(tan−1 α) − ⋅ 2α tan− 1 α d α +
∫ −1
(tan α) d α + C
2 1+ α 2 1 + α2
Partial Derivatives and their Applications 351
1
F(α) = log(1 + α2 ) ⋅ (tan−1 α) + C …(4)
2
At α = 0; from (1), F(0) = 0
implying C = 0
from(5), F(0) = 0 + C
α log(1 + αx)
∫
1
Hence dx = log(1 + α2 ) ⋅ (tan−1 α) …(5)
0 (1 + x2 ) 2
1 log(1 +αx) π
∫
1
Putting α = 1 in (5), dx = log 2(tan−1 1) = loge 2
0 (1 + x ) 2
2 8
a2
∫
d x 1
Example 91: Show that tan−1 dx = 2a tan−1 a − log (a2 + 1) .
da 0 a 2
tan −1 x dx
a2
Solution: Let F(a) = ∫0 a
so that
a2
∂ x a2 d 0
∫
d d
F(a) = tan−1 dx + (a2 )tan−1 − (0) ⋅ tan−1
da 0 ∂a a da a da a
(By Leibnitz’s Rule II.)
a2
d x −1 a2
∫ ∫
1 2x
= dx + 2a tan− a = dx + 2a tan−1a
1
2
a2 + x2
1 + x
0 da a 2 0
a
1 a2 1
= − log(a2 + x2 ) 0 + 2a tan−1 a = 2a tan−1 a − log(1 + a2 ) …(3)
2 2
x
Verification: Put = tan θ so that dx = a sec2θ dθ and limits for x = 0, θ = 0,
a
−1
for x = a , θ = tan a
2
a2 tan −1 a tan −1 a
∫ ∫ ∫
x tan −1 a
∴ tan−1 dx = a θ ⋅ sec2 θ dθ = a (θ ⋅ tan θ)0 − 1 ⋅ tan θ dθ …(4)
0 a 0 0
( )
tan−1 a tan−1 a
∫ tan θ dθ = ( log cos θ )0 = ( − log sec θ )0
tan−1 a tan−1 a
whereas = − log 1 + tan2 θ
0 0
(log(1 + tan2 θ))0
1 tan−1 a 1
=− = − log(1 + a2 ) …(5)
2 2
a2
∫
x a
implying tan−1 dx = a2 tan−1 a − log(1 + a2 ) …(6)
0 a 2
Differentiation of (6) with respect to ‘a’ gives the desired result.
352 Engineering Mathematics through Applications
∫
x
Example 92: If y = f (t)sin k(x − t)dt , prove that y satisfies the differential equation.
0
d2 y
+ k2 y = k f (x ).
dx2
∫
x
Solution: Given y= f (t)sin k(x − t) dt , x is parameter …(1)
0
By Leibnitz’s Rule 2,
∫
dy x
d
= f (t)cos k(x − t) ⋅ k dt + (x) f (x) ⋅ sin k(x − x) − 0
dx 0 dx
∫
x
=k f (t)cos k(x − t) dt + 0 …(2)
0
d2 y
∫
x
d dy d
∴ 2
= =k f (t) − sin k( x − t) k dt + k (x) f (x) ⋅ cos k(x − x) − 0
dx dx dx 0 dx
∫
x
= −k 2 f (t)sin k(x − t) dt + k ⋅ 1 ⋅ f (x)
0
d2 y
= −k2 y + k f (x)
dx2
d2 y
Hence + k 2 y = k f (x)
dx2
ASSIGNMENT 10
xα − 1
∫
1
1. Evaluate dx, α ≥ 0 [MDU 2004; KUK, 2004]
0 log x
2. By differentiating under the integral sign or otherwise show that
π log(1 + sin α ⋅ cos x)
∫
x
dx = πα Hint : Take tan = t [NIT Kurukshetra, 2007]
0 cos x 2
∞ − ax
∫
e sin x
3. By differentiating under the integral sign, evaluate the integral dx
0 x
∞
π
∫
sin x
Hence show that dx =
0 x 2
α2
tan−1 x dx
4. Verify Leibnitz rule for differentiation under integral sign if F(α) = ∫ 0 α
∫
1
∫ x (log x) dx
1 1
5. By successive differentiation of xm dx = with respect to m, evaluate m n
0 m+1 0
Partial Derivatives and their Applications 353
ANSWERS
Assignment 1
5. (1 + 3xyz + x2y2z2) exyz 8. 2
Assignment 3
x
6. – 7 tanu 8. 6x4 y2 sin−1
y
Assignment 4
log(sin y) + y tan x
1. 8 e4t 2. log(cos x) − x cot y
a2 x2 + y
3. 2 1 − 2 x cos(x + y ) 4. 1 + log xy − x
2 2
b y2 + x
32 y 1 y 1
5. − cm/sec. 7. 2yz − x , 4 ; y2 − 3z2 , − 11
21
Assignment 6
1. 4(u2 + v2) 2. x(u – 1 – vy) + 2uv – z
3. – 4r3; r2 = x2 + y2 4. 0; sin u = v
u
5. w2 = 2u + v2 6. w = 4 (u2 + 3v)
−1
7. (x − y)(y − z)(z − x) J' = (x − y)(y − z)(z − x)
Assignment 7
1. 2% 5. (p – 3q – 4r)%
4
6. − % 7. 5r
3
12. – 1
Assignment 8
1. ex cos y = e + (x − 1) e + y − π − e
2 2 4 2
354 Engineering Mathematics through Applications
1 2 e y − π − e + y − π − e + …
2
+ (x − 1) + 2(x − 1)
2 2 4 2 4 2
Hint : For a = 1, b = π ; x − a = h = x − 1, y − b = k = y − π
4 4
2
π2 π π π
(x − 1)2 − (x − 1) y − − y −
1
2. 1 −
8 2 2 2 2
1 2 1 2 1
3. 2.306 4. xy 5. y + xy − y + (x y − xy2 ) + y3 + …
2 2 3
Assignment 9
1. (i) Min at (a, a) (ii) Max at (± 1, 0); Min at (0, ± 1)
2. p2/(a2 + b2 + c2) 3. x = y = 2z
log k 2 log k 2 log k 2
bc ca ab
a b c
4. x = ,y= ,z=
3 log a 3 log b 3 log c
mm nn pp am + n + p
5. x = 12, y = 8, z = 4 6. (m + n + p)m + n + p 7. (0, 0, ±1)
x1 + x2 + x3 y +y +y
8. 50 9. x= , y= 1 2 3
3 3
−2r2 π
12. Areamaxi = at A = B = C =
3 3
Assignment 10
(−1)n n
1. [log(1 + α)] 5. ( n +1
m + 1)
You might also like
- Ged 102 Mathematics in The Modern World Module PDF83% (261)Ged 102 Mathematics in The Modern World Module PDF374 pages
- Mathematics in The Modern World 1st Year 1st Semester97% (91)Mathematics in The Modern World 1st Year 1st Semester46 pages
- Hist 1 Readings in Philippine History Module PDF86% (100)Hist 1 Readings in Philippine History Module PDF92 pages
- Differential and Integral Calculus: Sixth Edition80% (5)Differential and Integral Calculus: Sixth Edition608 pages
- Midterm Exam in Statistics and Probability (Grade 11)95% (59)Midterm Exam in Statistics and Probability (Grade 11)2 pages
- STATISTICS AND PROBABILITY Reviewer With Answer and Solutions100% (5)STATISTICS AND PROBABILITY Reviewer With Answer and Solutions14 pages
- 40 Multiple Choice Questions in Basic Statistics89% (9)40 Multiple Choice Questions in Basic Statistics8 pages
- Mathematics-in-the-Modern-World-Module by CHED90% (10)Mathematics-in-the-Modern-World-Module by CHED444 pages
- Mathematics - 120 Questions With Answer Key - LET EXAM - Questions & Answers100% (5)Mathematics - 120 Questions With Answer Key - LET EXAM - Questions & Answers19 pages
- Statistics-and-Probability - G11 - Quarter - 4 - Module - 1 - Test-of-Hypothesis85% (87)Statistics-and-Probability - G11 - Quarter - 4 - Module - 1 - Test-of-Hypothesis35 pages
- Statistics and Probablity SHS 11-Module 1 - Week194% (47)Statistics and Probablity SHS 11-Module 1 - Week123 pages
- MCQ Measures of Central Tendency With Correct Answers PDF91% (11)MCQ Measures of Central Tendency With Correct Answers PDF10 pages
- MODULE in Mathematics in The Modern World Unit 169% (16)MODULE in Mathematics in The Modern World Unit 143 pages
- Finite Differences and Interpolation, Numerical Differentiation and Integration100% (2)Finite Differences and Interpolation, Numerical Differentiation and Integration92 pages
- Mathematical Methods in Chemical Engineering87% (15)Mathematical Methods in Chemical Engineering576 pages
- Engineering Mathematics I Question Bank With Answer KeyNo ratings yetEngineering Mathematics I Question Bank With Answer Key10 pages
- Introduction To CFX: Interfaces, Sources and Additional VariablesNo ratings yetIntroduction To CFX: Interfaces, Sources and Additional Variables25 pages
- Conjugate Harmonic Function Harmonic Conjugate Function: Milne - Thomson's Method100% (3)Conjugate Harmonic Function Harmonic Conjugate Function: Milne - Thomson's Method5 pages
- Equations Reducible To Variables SeparableNo ratings yetEquations Reducible To Variables Separable4 pages
- Homogeneous Linear Partial Differential Equations With Constant CofficentsNo ratings yetHomogeneous Linear Partial Differential Equations With Constant Cofficents17 pages
- Double Integrals and Change of Order of Integration100% (4)Double Integrals and Change of Order of Integration12 pages
- 6.homogeneous Linear Equations With Constant Coefficients100% (1)6.homogeneous Linear Equations With Constant Coefficients26 pages
- Differential Equations and Vector CalculusNo ratings yetDifferential Equations and Vector Calculus361 pages
- Linear Differential Eqations of Second and Higher Order100% (1)Linear Differential Eqations of Second and Higher Order52 pages
- Linear Differential Equations With Constant Coefficient75% (4)Linear Differential Equations With Constant Coefficient9 pages
- Linear Partial Differential Equations of High Order With Constant CoefficientsNo ratings yetLinear Partial Differential Equations of High Order With Constant Coefficients58 pages
- Limits and Continuity of Functions of Several Variables PDF100% (1)Limits and Continuity of Functions of Several Variables PDF28 pages
- Module 2 - Partial Derivatives - PresentationNo ratings yetModule 2 - Partial Derivatives - Presentation22 pages
- Mathematics B3D Dr. Helen J. Wilson Spring 2006No ratings yetMathematics B3D Dr. Helen J. Wilson Spring 200677 pages
- Partial Differential Equations Arise in Connection With Various Physical andNo ratings yetPartial Differential Equations Arise in Connection With Various Physical and99 pages
- Unit-3 Multivariable Fucntion and partial differentiation(theory)No ratings yetUnit-3 Multivariable Fucntion and partial differentiation(theory)36 pages
- Problem-Solving, Mathematical Investigation and Modelling: MATH 110100% (2)Problem-Solving, Mathematical Investigation and Modelling: MATH 11035 pages
- Most Essential Learning Module in Mathematics 2020No ratings yetMost Essential Learning Module in Mathematics 202014 pages
- IEO Syllabus Third Year-Computer Science EngineeringNo ratings yetIEO Syllabus Third Year-Computer Science Engineering2 pages
- Material Selection in Mechanical DesignNo ratings yetMaterial Selection in Mechanical Design29 pages
- LINEAR PROGRAMMING - Simplex MinimizationNo ratings yetLINEAR PROGRAMMING - Simplex Minimization6 pages
- Development of Design Criteria For CCTV Camera PolesNo ratings yetDevelopment of Design Criteria For CCTV Camera Poles4 pages
- Math9 q1 Wk1 Mod1 Illustrating-Quadratic-EquationsNo ratings yetMath9 q1 Wk1 Mod1 Illustrating-Quadratic-Equations14 pages
- Florante D. Cenas, LPT, MAMT Abigail V. Santisteban, LPT Baby Jean Baligasa, LPT100% (1)Florante D. Cenas, LPT, MAMT Abigail V. Santisteban, LPT Baby Jean Baligasa, LPT62 pages
- Independent and Dependent Variables: Continuous and Discrete Data CorrelationNo ratings yetIndependent and Dependent Variables: Continuous and Discrete Data Correlation22 pages
- Linear Equations in Two Variables: Table of ContentsNo ratings yetLinear Equations in Two Variables: Table of Contents14 pages
