MATHunit 1 (2) 2

TRANSFORM AND BOUNDARY VALUE PROBLEMS
Dr.S.K.THAMILVANAN (21MAB201T) 1 / 96
Introduction UNIT-I
A partial Differential Equation is an equation involving a function of two

or more variables and some of its partial derivatives. Therefore a PDE
contains one dependent variable and more than one independent variable.
Hence the main difference between partial and ordinary differential
equation is the number of independent variables involved in the equation.
The order of a PDE is the order of the highest partial derivatives occurring
in the equation.
FORMATION OF PARTIAL DIFFERENTIAL EQUATIONS:
There are two methods to form a PDE.
1. By Elimination of arbitrary constants.
2. By Elimination of arbitrary functions.
Introduction UNIT-I
1. Formation of PDE by elimination of arbitrary constants:

Let f (x, y , z, a, b)=0 —-(1) be an equation which contains two arbitrary
constants a and b. To eliminate this two constant we need atleast three
equation. Therefore partially differentiating eqn (1) w.r.to x and y, we get
two more eqn.
From these three eqn we can eliminate a and b. Similarly , for eliminating
three constants we need four equations and so on.
We use the following notations
∂z
∂x = p
∂z
∂y =q
∂2z ∂2z ∂2z
∂x 2
= r, ∂y 2
= t, ∂x∂y = s.
Differential Equations in Civil Engineering
Differential equation of axial deformation of bar:

d
[A(x)σ(x)] + f (x) = 0,
dx
where σ(x) = deformation, A(x)=cross sectional area and f (x) =
force per unit length.
Differential equation of a elastic beam:
d2 d 2Y

EI = q,
dx 2 dx 2
where E = modulus of elasticity of the beam, I =moment of inertia of

the cross section and q = external load.
Differential Equations in Computer Science Engineering
Lagrange multipliers are always used to optimize and maximize the
functions of neural networks (AI-artificial intelligence) :
Gradient decent methods are used in back propagation for fitting an

artificial neural network:
Almost all of machine learning and AI (svm ,regressions trees) are

filled with examples of differential equations:
In computer graphics, animation, computer vision, and shape retrieval

such as denoising, fairing, feature extraction, segmentation,
simplification, classification, and editing; the complex manipulations
can be described precisely, compactly, and measurably by using PDEs:
If you develop a computer programme to predict future values,

especially on engineering quantities, you will almost certainly use
differential equations as part of your predictive algorithms in your tool
kit :
What is Partial Differential Equations(PDEs)? Unit-I
Definition: The differential equation involving more than one independent

variables and any number of dependent variables is called PDE.
Let z = f (x, y ) then
∂z
∂x = p = partial derivative of z w.r.t x
∂z
∂y = q = partial derivative of z w.r.t y
∂2z ∂2z ∂2z
∂x 2
= r, ∂y 2
= t, ∂x∂y = s.
2
∂z ∂z ∂z 2 ∂2z ∂z

Example: ∂x + ∂y = 1, ∂x +2 ∂x∂y + ∂y = 0 and so on....
Order and Degree of a PDE Unit-I
Order: The order of the PDE is the order of the highest order derivative
occurring in it.
Degree: The degree of the PDE is the power of its highest derivative
occurring in it.
∂z ∂z
Example: (i) ∂x + ∂y = 1, order=1, degree=1.
2
∂z 2 ∂2z ∂z

(ii) ∂x +2 ∂x∂y + ∂y = 0, order=2, degree=1.
3 3
∂2z 2 ∂z ∂2z ∂z 2

(iii) ∂x 2
= ∂x ⇒ ∂x 2
= ∂x , order=2, degree=3.
Formation of PDEs Unit-I
Let
f (x, y , z, a, b) = 0 (1)
be the given function, where z = z(x, y ) is the dependent variable depends

on x and y , a and b are arbitrary constants. Then we can form a PDE
from equation (1) in two ways:
Eliminating arbitrary constants
Eliminating arbitrary function
Elimination of arbitrary constants Unit-I
Let us consider equation (1) e.i. f (x, y , z, a, b) = 0 and we proceed as

below:
Step-I: Differentiate equation (1) w.r.t x
∂f ∂f ∂z ∂f ∂f
+ =0= +p (2)
∂x ∂z ∂x ∂x ∂z
Step-II: Differentiate equation (1) w.r.t y
∂f ∂f ∂z ∂f ∂f
+ =0= +q (3)
∂y ∂z ∂y ∂x ∂z
Eliminate the constants a and b from equations (2) and (3).
Finally, write the PDE in the form
ϕ(x, y , z, p, q) = 0. (4)
Note: The equation (1) is called the primitive or complete solution of the
equation (4).
Example 1: Form the PDE from z = (x 2 + a)(y 2 + b).
Solution: Differentiating w.r.t x we get

∂z
= 2x(y 2 + b) = p (5)
∂x
Differentiating w.r.t y we get
∂z
= 2y (x 2 + a) = q (6)
∂y
Equation(5) and (6) gives

p q
(y 2 + b) = and (x 2 + a) = respectively. (7)
2x 2y
Substituting equation (7) in the original equation

p q
z = (x 2 + a)(y 2 + b) = 2x . 2y ⇒ pq = 4zxy .
Example 2: Find the PDE of all planes cutting equal intercepts from the
x-axis and y -axis.
Solution: Let a and c are x and z intercepts respectively. The the

equation of plane is given by
x y z
+ + = 1. (8)
a a c
Differentiating w.r.t x we get
1 1 ∂z 1 1
+ = 0 = + p. (9)
a c ∂x a c
Differentiating w.r.t y we get
1 1 ∂z 1 1
+ = 0 = + q. (10)
a c ∂y a c
Solving (9) and (10) gives
1 1 1
p − q = 0 ⇒ (p − q) = 0 ⇒ p − q = 0.
c c c
Example 3: Form the PDE from log(az − 1) = x + ay + b.
Solution:
a ∂z ap a ∂z aq
=1= and =a= (11)
az − 1 ∂x az − 1 az − 1 ∂y az − 1
a ∂z aq
=a= (12)
az − 1 ∂y az − 1
q
Diving (12) by (11), we get p = a. Fron equation (12) we have
q
q = az − 1 = z − 1 ⇒ pq = qz − p ⇒ p(q + 1) = qz.
p
Example 4: Obtain the PDE of all spheres with centers lies on z = 0 and
whose radius is constant and equal to r .
Solution: The equation of the spheres with z = 0 is

(x − a)2 + (y − b)2 + z 2 = r 2 .
∂z
2(x − a) + 2z = 0 = 2(x − a) + 2zp and
∂x
∂z
2(y − b) + 2z = 0 = 2(y − b) + 2zq
∂y
⇒ (x − a) = −zp and (y − b) = −zq
⇒ z 2 p 2 + z 2 q 2 + z 2 = r 2 = z 2 (p 2 + q 2 + 1).
2
x2 2
Example 5: Form a PDE by eliminating a, b and c from a2
+ yb2 + cz 2 = 1.
Solution:
2x 2z ∂z ∂z
+ 2 =0 or c 2 x + a2 z =0 (13)
a2 c ∂x ∂x
2y 2z ∂z ∂z
2
+ 2 =0 or c 2 y + b2 z =0 (14)
b c ∂y ∂y
Again differentiating (13) w.r.t y we get
∂2z

2 ∂z ∂z
0+a z + =0 ⇒ pq + zs = 0. (15)
∂x∂y ∂x ∂y
Problems for practice: Unit-I
Problems for practice.
Find the PDE of the set of all right circular cones whose axes
coincides with z−axis and given by x 2 + y 2 = (z − c)2 tan2 α.
Ans: yp = xq
x2 y2
Form a PDE by eliminating a and b, from 2z = a2
+ b2
Ans: 2z = px + qy
Form a PDE by eliminating a and b, from 2z = (ax + y )2 + b
Ans: q = qp x + y
Elimination of arbitrary functions Unit-I
In order to form a PDE by eliminating arbitrary functions, we may come

across two types of functions
Form I: z = f (x, y ), where x and y are independent variables.
Form II: ϕ(u, v ) = 0, where u = u(x, y , z), v = v (x, y , z) and x, y

and z are independent variables.
Form I Unit-I
Example 1: Form the PDE by eliminating the arbitrary function from

z = f (x 2 + y 2 ).
Solution: Differentiating w.r.t x and y separately, we get

∂z ′ ∂z ′
= 2xf (x 2 + y 2 ) = p and = 2yf (x 2 + y 2 ) = q.
∂x ∂y
Dividing them we get
p x
= .
q y
Form I Unit-I

z = f (x − at) + f (x + at).
Solution: Differentiating w.r.t x, we get

∂z ′ ′
= f (x − at) + f (x + at)
∂x
∂2z ′′ ′′
= f (x − at) + f (x + at). (16)
∂x 2
Similarly differentiating w.r.t t, we get
∂z ′ ′
= −af (x − at) + af (x + at)
∂t
∂2z ′′ ′′
= a2 f (x − at) + a2 f (x + at). (17)
∂t 2
Conti..... Unit-I
Form (17) one can see that
∂2z ′′ ′′
= a2 f (x − at) + a2 f (x + at)
∂t 2
′′ ′′
= a2 [f (x − at) + f (x + at)]
2
∂ z
= a2 . From (16)
∂x 2
∂2z 2
∂ z
Therefore the required PDE is ∂t 2
= a2 ∂x 2.
Form I Unit-I
y = f yx .
∂z ′
y −y ′
y −x 2 ∂z
=f or f = (18)
∂x x x2 x y ∂x
and

∂z ′
y 1 ′
y ∂z
=f or f =x (19)
∂y x x x ∂y
From (18) and (19) one can find

2
−x ∂z ∂z ∂z ∂z
=x ⇒x +y = 0.
y ∂x ∂y ∂x ∂y
Form I Unit-I
Example 4: Form the PDE

by eliminating the arbitrary function from
2 1
y = x + 2f y + log x .

∂z ′ 1 1
= p = 2x + 2f + log x and
∂x y x

∂z ′ 1 −1
= q = 2f + log x
∂y y y2
⇒ px + qy 2 = 2x 2
p − 2x y2
⇒ =− .
q x
Form I Unit-I

y = xy + 2f (x 2 + y 2 + z 2 ).

∂z ′ 2 2 2 ∂z
= p = y + 2f (x + y + z ) 2x + 2z and
∂x ∂x

∂z ′ 2 2 2 ∂z
= q = x + 2f (x + y + z ) 2y + 2z
∂y ∂y
′
⇒ p − y = 2f (x 2 + y 2 + z 2 ) (2x + 2zp)
′
⇒ q − x = 2f (x 2 + y 2 + z 2 ) (2y + 2zq)
p−y x + zp
⇒ = .
q−x y + zq
Problems for practice: Unit-I
Problems for practice.
Form the PDE by eliminating the arbitrary functions f and g

z = f (ax + by ) + g (αx + βy ).
∂ z 2 ∂z 2
∂ z
Ans: bβ ∂x 2 − (aβ + bα) ∂x∂y + aα ∂x 2 = 0
Form a PDE by eliminating f and g , from z = yf (x) + xg (y )

∂z ∂z ∂z
Ans: xy ∂x∂y = x ∂x + y ∂y −z
Eliminate the arbitrary function f and obtain the PDE
from z = e y f (x + y ).
∂z ∂z
Ans: ∂y =z+ ∂x
Form II Unit-I
Let ϕ(u, v ) = 0 be given function. Then we can construct the PDE as

follows:
Step 1: Differentiate u and v w.r.t x, y and z
Step 2: Find
∂u ∂v ∂u ∂v ∂(u, v )
P= − =
∂y ∂z ∂z ∂y ∂(y , z)
∂u ∂v ∂u ∂v ∂(u, v )
Q= − =
∂z ∂x ∂x ∂z ∂(z, x)
∂u ∂v ∂u ∂v ∂(u, v )
R= − =
∂x ∂y ∂y ∂x ∂(x, y )
Step 3: Write the PDE Pp + Qq = R.
Form II Unit-I
Example 1: Form the PDE by eliminating the arbitrary function ϕ from
ϕ(x + y + z, x 2 + y 2 − z 2 ) = 0.
Solution: Here u = x + y + z and v = x 2 + y 2 − z 2

∂u ∂u ∂u
= 1, = 1, =1
∂x ∂y ∂z
∂v ∂v ∂v
= 2x, = 2y , = −2z.
∂x ∂y ∂z
Now
∂u ∂v ∂u ∂v
P= − = 1.(−2z) − 1.(2y ) = −2z − 2y
∂y ∂z ∂z ∂y
∂u ∂v ∂u ∂v
Q= − = 1.(2x) − 1.(−2z) = 2x + 2z
∂z ∂x ∂x ∂z
∂u ∂v ∂u ∂v
R= − = 1.(2y ) − 1.(2x) = 2y − 2x.
∂x ∂y ∂y ∂x
Conti..... Unit-I
Write the PDE as
Pp + Qq = R
⇒ (−2z − 2y )p + (2x + 2z)q = 2y − 2x
= (y + z)p − (x + z)q = x − y
Form II Unit-I

ϕ(x 2 + y 2 + z 2 , z 2 − 2xy ) = 0.
Solution: Here u = x 2 + y 2 + z 2 and v = z 2 − 2xy

∂u ∂u ∂u
= 2x, = 2y , = 2z
∂x ∂y ∂z
∂v ∂v ∂v
= −2y , = −2x, = 2z.
∂x ∂y ∂z
P = (2y )(2z) − (2z)(−2x) = 4z(y + x)

Q = (2z)(−2y ) − (2x)(2z) = −4z(y + x)
R = (2x)(−2x) − (2y )(−2y ) = 4(y 2 − x 2 ).
y −x
Hence Pp + Qq = R ⇒ p − q = z .
Form II Unit-I
Example
3: Form the PDE by eliminating the arbitrary function f from
x−a y −b
f z−c , z−c = 0.
x−a y −b
Solution: Here u = z−c and v = z−c
∂u 1 ∂u ∂u x −a
= , = 0, =−
∂x z −c ∂y ∂z (z − c)2
∂v ∂v 1 ∂v y −b
= 0, = , =− .
∂x ∂y z −c ∂z (z − c)2
Conti...... Unit-I

y −b x −a 1 x −a
⇒ P = (0) − 2
− − 2
=
(z − c) (z − c) z −c (z − c)3

x −a 1 y −b y −b
Q = − (0) − − =
(z − c)2 z −c (z − c)2 (z − c)3
1 1 1
R = . − (0)(0) = .
z −c z −c (z − c)2
Hence the PDE is
Pp + Qq = R
(x − a)p + (y − b)q = (z − c).
Form II Unit-I

the relation ϕ(x 2 + y 2 + z 2 , lx + my + nz) = 0.
Solution: Here u = x 2 + y 2 + z 2 and v = lx + my + nz

∂u ∂v ∂u ∂v
P = − = (2y ).n − (2z).m
∂y ∂z ∂z ∂y
∂u ∂v ∂u ∂v
Q = − = (2z).l − (2x).m
∂z ∂x ∂x ∂z
∂u ∂v ∂u ∂v
R = − = (2x).m − (2y ).n.
∂x ∂y ∂y ∂x
Pp + Qq = R ⇒ 2(ny − mz)p + 2(lz − nx)q = 2(mx − ly )

= (ny − mz)p + (lz − nx)q = (mx − ly ).
Solutions of First Order Non-Linear PDEs Unit-I
Solution (Integral): The solution of a PDE is a relation between the

independent variables and dependent variables which satisfies the given
PDE.
There are four types of solutions:

Complete Integral
Particular Integral
Singular Integral
General Integral.
Complete Integral Unit-I
Definition: A solution which contains as many arbitrary constants as

there are independent variables is complete integral.
Example: If f (x, y , z, p, q) = 0 where x and y are independent variables is

a given PDE, then the solution of the form ϕ(x, y , z, a, b) = 0 is a
complete integral.
Particular Integral Unit-I
Definition: Let ϕ(x, y , z, a, b) = 0 be a complete integral of the given

PDE where a and b are arbitrary constants. If we give particular values to
a and b, we get particular integral.
Example: If ϕ(x, y , z, a, b) = 0 is a complete integral, then

ϕ(x, y , z, 1, 2) = 0, ϕ(x, y , z, 0, 1) = 0, ϕ(x, y , z, 3, 9) = 0 all are possible
complete integral.
Singular Integral Unit-I
Definition: Let
ϕ(x, y , z, a, b) = 0 (20)
be a complete integral of a given PDE. Differentiating (20) w.r.t a and b,

one can obtain
∂ϕ ∂ϕ
=0 and = 0. (21)
∂a ∂b
Now eliminating a and b from (20) and (21) we get singular integral.
General Integral Unit-I
Definition: Let ϕ(x, y , z, a, b) = 0 be a complete integral of

f (x, y , z, p, q) = 0. If we take b = φ(a) or a = φ(b) we get
ϕ(x, y , z, a, φ(a)) = 0 or ϕ(x, y , z, φ(b), b) = 0 (22)
Differentiating (22) w.r.t a we can obtain
∂ϕ
= 0. (23)
∂a
Now eliminating a from (22) and (23) we get general integral.
Types of First Order Non-Linear PDEs Unit-I
In order to find the solutions of some first order non-linear PDEs, let us
classify the given PDE as below:
Type I
Type II (Clairaut’s form)
Type III
Type IV (Separable equations).
Type I: Unit-I
Let us consider the first order PDEs of the form f (x, y , z, p, q) = 0

∂z ∂z
where p = ∂x and q = ∂y
Type I: If F (p, q) = 0, i.e it contains only p and q then the solution can
be obtained as below:
Step 1 : Put p = a and q = b which gives f (a, b) = 0

Step 2 : Assume the solution as z = ax + by + c
Step 3 : From f (a, b) = 0, write a = φ(b) or b = φ(a)
Step 4 (Complete Integral): Write the complete integral as
z = ax + φ(a)y + c (24)
Type I: Unit-I
Step 5 (Singular solution): Obtain the singular solution by

eliminating a and c from the equations (24) and
∂z ∂z
= 0, =0
∂a ∂b
Step 6 (General solution): Assume c = ψ(a) and eliminate a from
the following equations
∂z ′ ′
z = ax + φ(a)y + ψ(a) and = 0 = x + φ (a)y + ψ (a).
∂a
Type I Unit-I
√ √
Example 1: Solve p+ q = 1.
Solution: Consider
√ √ √ √
F (p, q) = p+ q − 1 ⇒ F (a, b) = a + b − 1.
Write the complete integral as

√ √
z = ax + by + c with a+
b = 1.
√ √ √ 2
⇒ b =1− a ⇒ b = (1 − a)
Therefore, the complete integral is given by

√
z = ax + (1 − a)2 y + c, (25)
where a and c are arbitrary constants.
Type I Unit-I
For the singular integral, differentiate (25) w.r.t c and equate with 0, we
get
∂z
= 0, ⇒ 1 = 0, which is absurd.
∂c
Therefore, there is no singular solution. For the general solution take
c = ψ(a), hence we get
√
z = ax + (1 − a)2 y + ψ(a). (26)
Now differentiation of (26) w.r.t a gives
∂z √ −1 ′
= x + 2(1 − a) × √ + ψ (a) = 0. (27)
∂a 2 a
Eliminating a from (26) and (27), we get general solution.
Type I Unit-I
Example 1: Solve p 2 + q 2 = npq.
Solution: Consider
F (p, q) = p 2 + q 2 − npq = 0 ⇒ F (a, b) = a2 + b 2 − nab = 0

a h p i
⇒ b= n ± n2 − 4 .
2
Complete integral:
ah p i
z = ax + n ± n2 − 4 y + c.
2
Singular integral: Differentiating the above equation w.r.t c
∂z
= 0, ⇒ 1 = 0, which is absurd.
∂c
Therefore, there is no singular solution.
Type I Unit-I
For the general solution take c = ψ(a), hence we get

ah p i
z = ax + n ± n2 − 4 y + ψ(a). (28)
2
∂z 1 h p i ′
= x + × n ± n2 − 4 y + ψ (a) = 0. (29)
∂a 2
Type II(Clairaut’s form) Unit-I
For the general solution take c = ψ(a), hence we get

ah p i
z = ax + n ± n2 − 4 y + ψ(a). (30)
2
∂z 1 h p i ′
= x + × n ± n2 − 4 y + ψ (a) = 0. (31)
∂a 2
Type II(Clairaut’s form): Unit-I
Let the first order PDEs is of the form z = px + qy + f (p, q) where
∂z ∂z
p = ∂x and q = ∂y . then the solution can be obtained as below:
Complete integral: Put p = a and q = b and write the complete

integral
z = ax + by + f (a, b). (32)
Singular integral: Obtain the singular integral by eliminating a and b

from (32) and the following equations
∂f
Diff. w.t.t a ⇒x+ =0 and (33)
∂a
∂f
Diff. w.t.t a ⇒y+ = 0. (34)
∂b
Type II(Clairaut’s form): Unit-I
General integral: Putting b = ψ(a) in (32), we get
z = ax + ψ(a)y + f (a, ψ(a)). (35)
Differentiation of (35) w.r.t a gives
∂z ′ ′
= x + ψ (a)y + f (a) = 0. (36)
∂a
Eliminating a from (35) and (36), one can get the general integral.
Type II Unit-I
p
Example 1: Solve z = px + qy + 1 + p2 + q2.
Solution: p
This is of the form z = px + qy + f (p, q) where
f (p, q) = 1 + p 2 + q 2 . √
Complete integral: It is given by z = ax + by + 1 + a2 + b 2 .
p
z = ax + by + 1 + a2 + b 2 . (37)
Singular integral: Differentiate (37) w.r.t a and b and eliminate them as

follows:
a a
x+√ = 0 ⇒ x = −√
1 + a2 + b 2 1 + a2 + b 2
and
b b
y+√ = 0 ⇒ y = −√ . (38)
1 + a2 + b 2 1 + a2 + b 2
Type II Unit-I
From the above equations one can find
a2 b2
x2 + y2 = +
1 + a2 + b 2 1 + a2 + b 2
1
⇒ 1 − x2 − y2 =
1 + a2 + b 2
1 p 1
⇒ 1 + a2 + b 2 = 2 2
⇒ 1 + a2 + b 2 = p .
1−x −y 1 − x2 − y2
As a result (38) becomes

p p
x = −a 1 − x 2 − y 2 and y = −b 1 − x2 − y2
x y
⇒ a = −p and b = −p .
1 − x2 − y2 1 − x2 − y2
Type II Unit-I
Therefore the singular integral w.r.t the equation (37) is given by
x2 y2 1
z = −p −p +
1 − x2 − y2 1 − x2 − y2 1 − x2 − y2
1 − x2 − y2 p
z = p = 1 − x2 − y2
1 − x2 − y2
⇒ z2 = 1 − x2 − y2
.
z 2 +x 2 +y 2 =1
Type II Unit-I
Example 2: Solve z = px + qy + p 2 q 2 .
Solution: This is of the form z = px + qy + f (p, q) where f (p, q) = p 2 q 2 .

Complete integral: It is given by
z = ax + by + a2 b 2 . (39)

follows:
x + 2ab 2 = 0 ⇒ x = −2ab 2
and
2
y + 2a b = 0 ⇒ y = −2a2 b. (40)
Type II Unit-I
From the above equations one can find

x y 1
= = −2ab = ⇒ a = kx and b = ky . (41)
a b k
Using (41) in (41) we get
1
x = −2k 3 x 2 y ⇒ k3 = −
2xy

4 2 2 1 2 2 3
z = kxy + kxy + k x y = 2kxy + kx y − = kxy
2xy 2

27 3 3 3 27 3 3 3 1 27
⇒ z3 = k x y = k x y − = − x 2y 2
8 8 2xy 16
.
16z 3 + 27x 2 y 2 = 0
Type II Unit-I
Example 3: Solve z = px + qy + p 2 − q 2 .
Solution: This is of the form z = px + qy + f (p, q) where

f (p, q) = p 2 − q 2 .
Complete integral: It is given by
z = ax + by + a2 − b 2 . (42)

follows:
x + 2a = 0 ⇒ x = −2a and y − 2b = 0 ⇒ y = 2b.

x y x2 y2 x2 − y2
⇒a=− and b = − ⇒ z = − + +
2 2 2 2 4
.
4z = y2 − x2
Type II Unit-I
General integral: Put b = ψ(a) in (42) becomes
z = ax + ψ(a)y + a2 − (ψ(a))2 .
Differentiating the above equation w.r.t a we find

′ ′
0 = x + ψ (a)y + 2a − 2ψ(a)ψ (a).
Eliminating a from the above equations, we will get the general solution.
Type III:Case I Unit-I
The PDEs of the form F (z, p, q) = 0 which does not contain x and y is
called Type-III problems. Assume the trial solution u = x + ay and take
z
= f (x + by ) = f (u)
∂z dz ∂u dz ∂z dz ∂u dz
⇒p = = . = and q = = . =a .
∂x du ∂x du ∂y du ∂y du
Then find the solutions as follows:
dz dz
Step I: Putting p = du and q = a du with u = x + by , we will get the
dz
ODE of the form du = ϕ(z, a).
dz dz
Step II: Solve the ODE du = ϕ(z, a) as ϕ(z,a) = du. Integrating this
R dz
we get ϕ(z,a) = u + c = x + by + c, where c is arbitrary constant.
Step III: Obtain the singular and general integral as usual.
Example 1: Solve p(1 + q) = qz.
Solution: This is of the form Type III. Assume u = x + by and put

dz dz
p = du and q = a du in the given equation.
2
dz dz dz dz dz
1+a = az ⇒a + (1 − az) =0
du du du du du
dz dz
⇒a = (1 − az) ⇒ a = du
du (1 − az)
Z Z
dz
⇒a = du + c ⇒ log(1 − az) = x + ay + c.
(1 − az)
Therefore, the complete integral is

.
log(1 − az) = x + ay + c
Singular integral: Differentiate w.r.t a and c we get

z
=y 0 = 1, which is absurd.
(1 − az)
∴ Singular solution does not exist.

General integral: Put c = ψ(a) in log(1 − az) = x + ay + c differentiate
it w.r.t a
z ′
= y + ψ (a)
(1 − az)
and eliminating afrom the above we will get the general integral.
Example 2: Solve 9(p 2 z + q 2 ) = 4.
Solution: This is of the form Type III. Assume u = x + ay and put

dz dz
p = du and q = a du in the given equation.
" 2 #
dz 2
2
2 dz dz
9 z +a =4 ⇒9 (z + a2 ) = 4
du du du
2
dz 4 dz 2
⇒ = ⇒ = p
du 9(z + a2 ) du 3 (z + a2 )
q
⇒ 3 (z + a2 )dz = 2du integrating we get (z + a2 )3/2 = u + b
∴ The complete integral is given by
(z + a2 )3 = (x + ay + b)2
Type III:Case II Unit-I
Suppose the equation is of the form f (x, p, q) = 0, which contains x, p

and q. This equation is also Type III. To solve such equation we use the
following steps:
Step I: Assume q = a and write p = ϕ(x, a) from f (x, p, a) = 0

Step II: Put the values of p and q in dz = pdx + qdy which gives
dz = ϕ(x, a)dx + qdy
Step III: Integrate the above equation and write the complete
integral as z = ϕ1 (x, a) + ay + c, where c is integrating constant and
ϕ1 (x, a) is integration of ϕ(x, a) w.r.t x.
Type III:Case II Unit-I
Example 1: Solve q = px + p 2 .
Solution: This is of the form Type III.
p 2 + px − q = 0, put q = a ⇒ p 2 + px − a = 0.
The above
√ equation is a quadratic equation in term of p. Therefore,
−x± x 2 +4a
p= 2 = ϕ(x, a).
dz = pdx + qdy
√ " √ #
−x ± x 2 + 4a −x x 2 + 4a
= dx + ady = ± dx + ady .
2 2 2
Integrating this, we get
−x 2
h √ i
z= 4 ± 1
2 sin h−1 2√
x
a
+ x
2 x 2 + 4a + ay + b
Type III:Case III Unit-I
Suppose the equation is of the form f (y , p, q) = 0, which contains y , p

and q. This equation is also Type III. To solve such equation we use the
following steps:
Step I: Assume p = a and write q = ϕ(y , a) from f (y , p, a) = 0

Step II: Put the values of p and q in dz = pdx + qdy which gives
dz = adx + ϕ(y , a)dy
Step III: Integrate the above equation and write the complete
integral as z = ax + ϕ1 (y , a)y + c, where c is integrating constant
and ϕ1 (y , a) is integration of ϕ(y , a) w.r.t y .
Type III:Case III Unit-I
Example 1: Solve pq = y .
Solution: Put p = a ⇒ q = ya . Now

y
dz = pdx + qdy = adx + dy ⇒ adz = a2 dx + ydy .
a
Integrating this, we get
az = ax 2 + y 2 /2 + c
is the complete integral. There will be no singular integral and the general
integral can be obtain by taking b = ψ(a) and eliminating a.
Type IV (Separable Equations): Unit-I
The first order PDE is said to be separable equation it it can put in the
form f (x, p) = ϕ(y , p).
For such PDE the solutions can be obtained as:
Step I: Put f (x, p) = ϕ(y , p) = a

Step II: Write p = f1 (x, a) and q = ϕ1 (y , a)
Step III: Putting in dz = pdx + qdy and integrating, get the
complete integral as
Z Z
z = f1 (x, a)dx + ϕ1 (y , a)dy + b.
Type IV: Unit-I
Example 1: Solve p 2 y (1 + x 2 ) = qx 2 .
Solution: This is separable equation because

√
p 2 (1 + x 2 ) q 2 ax 2 ax
2
= =a ⇒p = 2
⇒p= √ and q = ay .
x y (1 + x ) 1 + x2
√
ax
dz = pdx + qdy = √ dx + aydy .
1 + x2
Integrating the above equation we get the complete integral as
ay 2
p
z= a(1 + x 2 ) + 2 +b
Type IV: Unit-I
Example 2: Solve p 2 + q 2 = x + y .

√
p2 − x = y − q2 = a ⇒ p2 = x + a ⇒ p = x +a and
2 √
q = y − a ⇒ q = y − a.
Now
√ √
dz = pdx + qdy = x + adx + y − ady .
z = 23 (x + a)3/2 + 23 (y − a)3/2 + b
Type IV: Unit-I
Example 3: Solve p 2 + q 2 = x 2 + y 2 .

p
p 2 − x 2 = y 2 − q 2 = a2 ⇒ p 2 = x 2 + a2 ⇒ p = x 2 + a2 and
p
q 2 = y 2 − a2 ⇒ q = y 2 − a2 .
Now
p p
dz = pdx + qdy = x 2 + a2 dx + y 2 − a2 dy .
a2
√ y a2
cos h−1 ya + b
p
z= 2 sin h−1 xa + x
2 x 2 + a2 + 2 y 2 − a2 − 2
Lagrange’s Linear Equation Unit-I
A linear first order PDE of the form
Pp + Qq = R, (43)
where P, Q, R are are functions of x, y , z is called the Lagrange’s linear

equation. This equation can be solved in two ways
Method I (Direct method):
Method II (Method of multipliers):
Method I: Unit-I
The following steps are followed to solve (43):
dx dy dz
Step 1: Form the auxiliary simultaneous equations P = Q = R
Step 2: Solve the auxiliary equation to get two independent solutions

u(x, y , z) = a and v (x, y , z) = b
Step 3: Write the solution as ϕ(u, v ) = 0, where ϕ is an arbitrary

function.
Method I: Unit-I
Example 1: Find the general solution of px + qy = z.
Solution: Write the auxiliary equation as

dx dy dz
= =
x y z
Consider the following pair of equations and solve them
dx dy dy dz
= and =
x y y z
x y
⇒ = u and = v.
y z
∴ The general solution is given by
ϕ( yx , yz ) = 0
Method I: Unit-I
y 2z
Example 2: Solve x p + xzq = y 2 .

xdx dy dz
2
= = 2
y z xz y
Consider the following pair of equations and solve them
xdx dy
2
= ⇒ x 2 dx = y 2 dy ⇒ x3 − y3 = u
y z xz
xdx dz
2
= 2 ⇒ xdx = ydy ⇒ x2 − y2 = v
y z y
ϕ(x 3 − y 3 , x 2 − y 2 ) = 0
Method I: Unit-I
Example 3: Solve tan xp + tan yq = tan z.

dx dy dz
= = ⇒ cot xdx = cot ydy = cot zdz
tan x tan y tan z
Taking cot xdx = cot ydy and integrating one can get
sinx
log sinx = log siny + log u ⇒ = u.
siny
Similarly taking
siny
cot ydy = cot zdz and integrating we get = v.
sinz

sinx siny
ϕ siny , sinz =0
Method II (Method of Multipliers): Unit-I
The following steps are followed to solve (43):

dx dy dz
Step 1: Write the auxiliary equations as P = Q = R
′ ′ ′
Step 2: Chose the multipliers l, m, n and l , m , n and rewrite the
′ ′ ′
ldx+mdy +ndz l dx+m dy +n dz
auxiliary equations as lP+mQ+nR = l ′ P+m′ Q+n′ R
such that
′ ′ ′
lP + mQ + nR = 0 and l P + m Q + n R = 0
Step 3: Then solve ldx + mdy + ndz = 0 and write the solution as
′ ′ ′
u = u(x, y , z). Similarly solve l dx + m dy + n dz = 0 and write the
solution as v = v (x, y , z).
Step 4: Write the general solution as ϕ(u, v ) = 0, where ϕ is an

arbitrary function.
Example 1: Solve (mz − ny )p + (nx − lz)q = (ly − mx).

dx dy dz
= =
(mz − ny ) (nx − lz) (ly − mx)
Consider the multipliers as x, y , z

xdx + ydy + zdz xdx + ydy + zdz
= .
x(mz − ny ) + y (nx − lz) + z(ly − mx) 0
Solving the following
xdx + ydy + zdz = 0 we get u(x, y , z) = x 2 + y 2 + z 2 .
Similarly considering the multipliers as l, m, n we have

ldx + mdy + ndz ldx + mdy + ndz
=
l(mz − ny ) + m(nx − lz) + n(ly − mx) 0
ldx + mdy + ndz = 0 we get v (x, y , z) = lx + my + nz.
ϕ(x 2 + y 2 + z 2 , lx + my + nz) = 0
Example 2: Solve x(z 2 − y 2 )p + y (x 2 − z 2 )q = z(y 2 − x 2 ).

dx dy dz
= =
x(z 2 − y 2 ) y (x 2 − z 2 ) z(y 2 − x 2 )
Consider the multipliers as x, y , z

xdx + ydy + zdz xdx + ydy + zdz
= .
x 2 (z 2 − y 2) 2 2 2 2 2 2
+ y (x − z ) + z (y − x ) 0
xdx + ydy + zdz = 0 we get u(x, y , z) = x 2 + y 2 + z 2 .
Similarly considering the multipliers as 1/x, 1/y , 1/z we have
1/xdx + 1/ydy + 1/zdz 1/xdx + 1/ydy + 1/zdz

= .
(z 2 2 2 2 2 2
− y ) + (x − z ) + (y − x ) 0
1/xdx + 1/ydy + 1/zdz = 0 we get v (x, y , z) = xyz.
ϕ(x 2 + y 2 + z 2 , xyz) = 0
Example 3: Solve (y + z)p + (z + x)q = (x + y ).

dx dy dz
= =
(y + z) (z + x) (x + y )
It can also be written as below:

dx + dy + dz dx − dy dy − dz
= = . Now taking
2(x + y + z) y −x z −y
dx + dy + dz dx − dy
= and integrating we get
2(x + y + z) −(x − y )
log(x + y + z) = −2 log(x − y ) + log u ⇒ u = (x + y + z)(x − y )2
Similarly considering
dx − dy dy − dz d(x − y ) d(y − z)
= ⇒ =
y −x z −y x −y y −z
and integrating we get
(x − y )
log(x − y ) = log(y − z) + log v ⇒v =
(y − z)

ϕ (x + y + z)(x − y )2 , (x−y )
(y −z) = 0
Example 4: Solve zp + yq = x.

dx dy dz
= =
z y x
Considering the following
dx dz
= and integrating we get
z x
u = x2 − z2
Similarly considering
dx + dy + dz dy d(x + y + z) dy
= ⇒ =
x +y +z y x +y +z y
and integrating we get
(x + y + z)
log(x + y + z) = log y + log v ⇒v =
y

ϕ x 2 − z 2 , (x+yy +z) = 0
PDEs of Higher Order Unit-I
Although PDEs of Higher Order can be classify in many categories but in

our course we will discuss only the following types of PDE:
Homogeneous Linear PDE:
Non-homogeneous Linear PDE:
Homogeneous Linear PDE: Unit-I
The PDE of the form

∂nz ∂nz ∂nz ∂nz
a0 + a1 + a 2 + .......... + +an = F (x, y ),
∂x n ∂x n−1 ∂y ∂x n−2 ∂y 2 ∂y n
where a0 , a1 , ........ an are the constants and F (x, y ) is the known

functions of x and y is known as homogeneous linear PDE with constants
co-efficient.
If we take ∂x∂
= D and ∂y ∂
= D ′ in the above equation, then it can be
written as

2 n
a0 D n + a1 D n−1 D ′ + a2 D n−2 D ′ + .......... + an D ′ z = F (x, y )
which can be further written as
f D, D ′ z = F (x, y ).

(44)
The solution of (44) consist of two parts:
Complementary function:
Particular Integral:
Complementary function(C.F): Unit-I
In order to find the complementary function the following steps may be

followed:
Step 1: Consider f (D, D ′ ) z = 0
Step 2: Put D = m and D ′ = 1 and write the auxiliary equation as
a0 mn + a1 mn−1 + a2 mn−2 + .......... + an = 0. (45)
Step 3: Solve the auxiliary equation (45) and find the roots. Let
m1 , m2 , ......mn are the n roots of this equation, then write the
complementary solution as per the following cases:
Complementary function(C.F): Unit-I
Case-I: If m1 ̸= m2 ̸= ..... ̸= mn i.e. that is all the roots are distinct, then
write the complementary function as:
z = ϕ1 (y +m1 x)+ϕ2 (y +m2 x)+ϕ3 (y +m3 x)+........+ϕn (y +mn x).
Case II: If m1 = m2 = ..... = mn i.e. all (may be some) roots are equal
(repeated) then write the complementary function as:
z = ϕ1 (y +m1 x)+xϕ2 (y +m1 x)+x 2 ϕ3 (y +m1 x)+....+x n−1 ϕn (y +m1 x).
Note: The degree of x may varies depending up on the repetition of the

particular root.
Particular Integral(P.I): Unit-I
Here we will discuss 6 possible types of functions to construct the P.I:
Type 1: If F (x, y ) = e ax+by , then P.I is given by
1 ax+by 1 ax+by ,
P.I = f (D, D ′ ) e = f (a, b) e f (a, b) ̸= 0
Note: Suppose f (a, b) = 0, then multiply the numerator by x and

differentiate the denominator by D, as a result we get
1 ax+by x n ax+by
f (D, D ′ ) e = n! e .
Type 2: If F (x, y ) = sin(ax + by ) or cos(ax + by ), then P.I is given by
1
P.I = sin(ax + by )/ cos(ax + by )
f D 2,DD ′ , D ′ 2
1
= sin(ax + by )/ cos(ax + by ),
f (−a , − ab, − b 2 )
2
provided f −a2 , − ab, − b 2 =

̸ 0.
Note: Suppose f −a2 , − ab, − b 2 = 0, then follow the previous note.

Type 3: If F (x, y ) = x r y s , then P.I is given by
1 r s
P.I = f (D, D ′ ) x y ,
where we can expand [f (D, D ′ )]−1 using binomial expansion in power of

D and D ′ .
Type 4: If F (x, y ) = ϕ(x, y )e ax+by , then P.I is given by
1 ax+by e ax+by
P.I = ϕ(x, y )e = ϕ(x, y ),
f (D, D ′ ) f (D + a, D ′ + b)
then proceed as type 3.
Type 5: If F (x, y ) = sin ax sin by , then P.I is given by
1 sin ax sin by
f −a2 , − b 2 =

P.I = . sin ax sin by = , ̸ 0.
f D 2, D ′2 f (−a2 , − b 2 )
Type 6: If F (x, y ) = cos ax cos by , then P.I is given by
1 cos ax cos by
f −a2 , − b 2 =

P.I = . cos ax cos by = , ̸ 0.
f D 2, D ′2 f (−a2 , − b 2 )
Note: If f −a2 , − b 2 = 0 in both type, then proceed as the note given

in Type 1.
∂3z 3 3
Example 1: Solve ∂x 3
− 3 ∂x∂2 ∂y
z ∂ z
+ 4 ∂y 3 = e
x+2y .
Solution: The above equation can be written as

3
D 3 − 3D 2 D ′ + 4D ′ z = e x+2y , here
3
f D, D ′ = D 3 − 3D 2 D ′ + 4D ′ and from e x+2y , a = 1, b = 2.

Put D = m and D ′ = 1 in the above equation and write the auxiliary

equation as
m3 − 3m2 + 4 = 0 ⇒ (m + 1)(m − 2)2 = 0 ⇒ m = −1, 2, 2,
where 2 is repeated root. Therefore, the C.F. is
C .F . = ϕ1 (y − x) + xϕ2 (y + 2x) + x 2 ϕ2 (y + 2x).
P.I.
1
P.I . = × e x+2y ,
D3 − 3D 2 D ′ + 4D ′ 3
1
⇒ × e x+2y
(1)3 − 3(1)2 (2) + 4(2)3
e x+2y
= .
27
∴ The complete solution is given by:
x+2y
C .F . + P.I . = ϕ1 (y − x) + xϕ2 (y + 2x) + x 2 ϕ3 (y + 2x) + e 27 .

Example 2: Solve D 3 − 7DD ′ 2 − 6D ′ 3 z = e 2x+y .
Solution: The auxiliary equation is
m3 − 7m − 6 = 0 ⇒ (m + 1)(m2 − m − 6) = 0
⇒ (m + 1)(m − 3)(m + 2) = 0 ⇒ m = −1, 3, − 2
∴ The C.F. is
C .F . = ϕ1 (y − x) + ϕ2 (y + 3x) + ϕ3 (y − 2x).
1 1
P.I . = ′ 2 ′ 3
× e 2x+y ⇒ 3 − 7(2)(1)2 − 6(1)3
× e 2x+y
D3 − 7DD − 6D (2)
e 2x+y
= − .
12
2x+y
z = C .F . + P.I . = ϕ1 (y − x) + ϕ2 (y + 3x) + ϕ3 (y − 2x) − e 12 .

Example 3: Solve D 2 − 4DD ′ + 4D ′ 2 z = e 2x+y .

m2 − 4m + 4 = 0 ⇒ (m − 2)2 ⇒ m = 2, 2. ∴ The C.F. is:
C .F . = ϕ1 (y + 2x) + xϕ2 (y + 2x).
1 1
P.I . = 2
× e 2x+y ⇒ × e 2x+y
D2 ′
− 4DD + 4D ′ (2) − 4(2)(1) + 4(1)2
2
e 2x+y
= − .
0
Hence we follow the Note.
Differentiate the denominator w.r.t D and multiply the numerator by x
x 2x+y e 2x+y
P.I . = × e =
2D − 4D ′ 0
∴ Once again differentiating the denominator w.r.t D and multiply the
numerator by x, we get
x2 e 2x+y
P.I . = × e 2x+y ⇒ .
2 2
∴ The complete integral is given by:
e 2x+y
z = C .F . + P.I . = ϕ1 (y + 2x) + xϕ2 (y + 2x) + 2 .
Example 4: Solve D 3 − 2D 2 D ′ z = sin(x + 2y ) + 3x 2 y .


m3 − 2m2 = 0 ⇒ m2 (m − 2) = 0 ⇒ m = 0, 0, 2.
∴ The C.F. is:
C .F . = ϕ1 (y ) + xϕ2 (y ) + ϕ2 (y + 2x).
Here we have to find two P.I. corresponding to two terms sin(x + 2y ) and
3x 2 y as in Type 2 and Type 3 respectively.
1 1
(P.I .)1 = ′
× sin(x + 2y ) and (P.I .)1 = 3 × 3x 2 y
D3 2
− 2D D D − 2D 2 D ′
1
(P.I .)1 = × sin(x + 2y )
D3 − 2D 2 D ′
1
= × sin(x + 2y )
D.D 2 − 2D.DD ′
1
= 2
× sin(x + 2y )
D. − (1) − 2D.(−1.2)
1 1
= × sin(x + 2y ) = × sin(x + 2y ).
−D + 4D 3D
1
As D is integration w.r.t x, hence:
(P.I .)1 = − 13 × cos(x + 2y ).
1 1
(P.I .)2 = ′
× 3x 2 y = 2
′ × 3x y
D3 − 2D D2
D 3 1 − 2D
D
′
−1
1 2D
= 3
1− × 3x 2 y (Using Binomial expansion)
D D
" #
1 2D ′ 4D ′ 2
= 1+ − + − − − − − − − × 3x 2 y .
D3 D D2
6x 2

1 2 1
= 3 3x 2 y + 2x 3

⇒ 3
3x y +
D D D
x 5y x6
Integrating w.r.t x thrice, we get (P.I .)2 = 20 + 60 . Finally we get
z = C .F . + (P.I .)1 + (P.I .)2
1 x 5y x6
= ϕ1 (y ) + xϕ2 (y ) + ϕ2 (y + 2x) − × cos(x + 2y ) + + .
3 20 60

MATHunit 1 (2) 2

Uploaded by

Copyright:

Available Formats

MATHunit 1 (2) 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MATHunit 1 (2) 2

Uploaded by

Copyright:

Available Formats

TRANSFORM AND BOUNDARY VALUE PROBLEMS

A partial Differential Equation is an equation involving a function of two

1. Formation of PDE by elimination of arbitrary constants:

Differential equation of axial deformation of bar:

where E = modulus of elasticity of the beam, I =moment of inertia of

Gradient decent methods are used in back propagation for fitting an

Almost all of machine learning and AI (svm ,regressions trees) are

In computer graphics, animation, computer vision, and shape retrieval

If you develop a computer programme to predict future values,

Definition: The differential equation involving more than one independent

be the given function, where z = z(x, y ) is the dependent variable depends

Eliminating arbitrary constants

Eliminating arbitrary function

Let us consider equation (1) e.i. f (x, y , z, a, b) = 0 and we proceed as

Step-I: Differentiate equation (1) w.r.t x

Eliminate the constants a and b from equations (2) and (3).

Finally, write the PDE in the form

Solution: Differentiating w.r.t x we get

Equation(5) and (6) gives

Substituting equation (7) in the original equation

Solution: Let a and c are x and z intercepts respectively. The the

Example 3: Form the PDE from log(az − 1) = x + ay + b.

Solution: The equation of the spheres with z = 0 is

Again differentiating (13) w.r.t y we get

Problems for practice.

In order to form a PDE by eliminating arbitrary functions, we may come

Form I: z = f (x, y ), where x and y are independent variables.

Form II: ϕ(u, v ) = 0, where u = u(x, y , z), v = v (x, y , z) and x, y

Example 1: Form the PDE by eliminating the arbitrary function from

Solution: Differentiating w.r.t x and y separately, we get

Example 2: Form the PDE by eliminating the arbitrary function from

Solution: Differentiating w.r.t x, we get

Form (17) one can see that

Solution: Differentiating w.r.t x and y separately, we get

From (18) and (19) one can find

Example 4: Form the PDE

Solution: Differentiating w.r.t x and y separately, we get

Example 5: Form the PDE by eliminating the arbitrary function from

Solution: Differentiating w.r.t x and y separately, we get

Problems for practice.

Form the PDE by eliminating the arbitrary functions f and g

Form a PDE by eliminating f and g , from z = yf (x) + xg (y )

Let ϕ(u, v ) = 0 be given function. Then we can construct the PDE as

Step 1: Differentiate u and v w.r.t x, y and z

Step 3: Write the PDE Pp + Qq = R.

Solution: Here u = x + y + z and v = x 2 + y 2 − z 2

Write the PDE as

Example 2: Form the PDE by eliminating the arbitrary function ϕ from

Solution: Here u = x 2 + y 2 + z 2 and v = z 2 − 2xy

P = (2y )(2z) − (2z)(−2x) = 4z(y + x)

Hence the PDE is

Example 4: Form the PDE by eliminating the arbitrary function ϕ from

Solution: Here u = x 2 + y 2 + z 2 and v = lx + my + nz

Pp + Qq = R ⇒ 2(ny − mz)p + 2(lz − nx)q = 2(mx − ly )

Solution (Integral): The solution of a PDE is a relation between the

There are four types of solutions:

Definition: A solution which contains as many arbitrary constants as

Example: If f (x, y , z, p, q) = 0 where x and y are independent variables is

Definition: Let ϕ(x, y , z, a, b) = 0 be a complete integral of the given

Example 4: Form the PDE