Unit-5

PARTIAL DIFFERENTIAL EQUATIONS

Introduction:
An equation involving partial differential coefficients of a function of two or more
variables is known as a partial differential equation. If a partial differential
equation contains nth and lower order derivatives, it is said to be of n th order PDE.
The degree of such equation is the greatest exponent of the highest order. Further
such equation will be called linear, if it is a first degree in the dependent variable
and its partial derivatives (i.e. powers or products of the dependent variable and its
partial derivatives must be absent). An equation which is not linear is called a non-
linear differential equation.
In case of two dependent variables we usually assume them to be x and y and z to
be dependent on x and y. If there are n-independent variables we take them to be
x1 , x2 , x3 ,……. xn and z is then regarded as the dependent variable.

Classification of 1st order partial differential equations:

Semi Linear Equations: A first order partial differential equation F(x, y, z, a, b)=0
is known as a semi linear equation if it is linear in a and b and the coefficients of a
and b are functions of x and y only.
Quasi linear Equations: A first order partial differential equation F(x, y, z, a, b)=0
is known as quasi linear partial differential equation if it is linear in a and b.
Non Linear Equations: A first order partial differential equation F(x, y, z, a, b)=0
which does not come under above types is known as nonlinear equation.

Examples:
𝜕2 𝑢 𝜕2 𝑢
• = is a wave equation , second order, first degree, linear , homogenous.
𝜕𝑡 2 𝜕𝑥 2
 𝜕𝑢 𝜕2 𝑢
• = is a heat equation , second order, linear, homogenous.
𝜕𝑡 𝜕𝑥 2
𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
• 2
+ 2
+ = 0 is a Laplace’s equation, second order , linear , homogenous.
𝜕𝑥 𝜕𝑦 𝜕𝑧 2

𝜕2 𝑢 𝜕2 𝑢
• 2
+ = 𝑓(𝑥, 𝑦) is a second order, linear and non-homogeneous.
𝜕𝑥 𝜕𝑦 2

• (𝑖) p2 + q2 = 1 (ii) pq = 𝑧 is a non-linear PDE.

The following are the standard notations when z = f (x, y) are used:
𝜕𝑧 𝜕𝑧 𝜕2 𝑧 𝜕2 𝑧 𝜕2 𝑧
=𝑝 =𝑞 =𝑟 =𝑠 =𝑡
𝜕𝑥 𝜕𝑦 𝜕𝑥 2 𝜕𝑥𝜕𝑦 𝜕𝑦 2

Solution of partial differential equations by direct method or direct

integration:
𝜕 𝑧 2 𝜕𝑧
1. Solve 𝜕𝑥𝜕𝑦 = 𝑠𝑖𝑛 𝑥 𝑠𝑖𝑛 𝑦 for which 𝜕𝑦 = −2 𝑠𝑖𝑛 𝑦when x=0 and z=0 when y is an
odd multiple of 𝜋2.
Solution:
𝜕2 𝑧
= 𝑠𝑖𝑛 𝑥 𝑠𝑖𝑛 𝑦
𝜕𝑥𝜕𝑦
Integrating with respect to x (keeping y constant)

𝜕𝑧
= − 𝑐𝑜𝑠 𝑥 𝑠𝑖𝑛 𝑦 + 𝑓(𝑦). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
𝜕𝑦
𝜕𝑧
When x=0 and = −2 𝑠𝑖𝑛 𝑦. Then we get f(y)= - siny
𝜕𝑦
Substituting the value of f(y) in equation (1),
𝜕𝑧
= − 𝑐𝑜𝑠 𝑥 𝑠𝑖𝑛 𝑦 − 𝑠𝑖𝑛 𝑦.
𝜕𝑦
Integrating again with respect to y (keeping x constant)

𝑧 = 𝑐𝑜𝑠 𝑥 𝑐𝑜𝑠 𝑦 + 𝑐𝑜𝑠 𝑦 + 𝑔(𝑥). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)

𝜋
When y is an odd multiple of at z=0. This implies g(x) =0
2

𝑧 = 𝑐𝑜𝑠 𝑥 𝑐𝑜𝑠 𝑦 + 𝑐𝑜𝑠 𝑦 = 𝑐𝑜𝑠 𝑦 (1 + 𝑐𝑜𝑠 𝑥)

 3
2. Solve 𝜕𝑥𝜕2𝜕𝑦
𝑧
= 𝑐𝑜𝑠( 2𝑥 + 3𝑦)
Solution:
𝜕3𝑧
= 𝑐𝑜𝑠( 2𝑥 + 3𝑦). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
𝜕𝑥 2 𝜕𝑦
Integrate eqn (1) wrt x (keeping y constant)

𝜕2𝑧 𝑠𝑖𝑛( 2𝑥 + 3𝑦)

= + 𝑓(𝑦). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
𝜕𝑥𝜕𝑦 2
Again integrating eqn (2) wrt x (keeping y constant)

𝜕𝑧 𝑐𝑜𝑠( 2𝑥 + 3𝑦)
=− + 𝑓(𝑦)𝑥 + 𝑔(𝑦). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
𝜕𝑦 4
Again integrating eqn (3) wrt y (keeping x constant)

𝑠𝑖𝑛( 2𝑥 + 3𝑦)
𝑧=− + 𝑥 ∫ 𝑓(𝑦)𝑑𝑦 + ∫ 𝑔(𝑦)𝑑𝑦 + ℎ(𝑥)
12

Linear Partial Differential Equations of 1st Order

The general form of a quasi-linear partial differential equation of 1st order is:
P(x,y,z) zx + Q(x,y,z)zy = R(x,y,z) ---------------(1)
This equation (1) is known as Lagrange’s linear equation.
If P and Q are independent of Z and R is linear in Z then (1) is a linear equation. The
general solution of Lagrange’s linear PDE:
Pp + Qq =R --------------- (1)
ss given by the equation;
F (u,v)=0 ---------------(2)
Since the elimination of arbitrary function F from (2) results in (1).
Here u=u(x,y,z), v=v(x,y,z) are known functions.
Method of obtaining general solution:
1) Rewrite the equation in the standard form Pp+Qq=R
2) Form the Lagrange’s auxiliary equation (A.E)
𝑑𝑥 𝑑𝑦 𝑑𝑧
= = --------------- (3)
𝑃 𝑄 𝑅

3) u(x,y,z)=C1 and v(x,y,z)=C2 are said to be the complete solution of the system
of simultaneous equations (provided u1 and u2 are linearly independent i.e
u1/u2 not equal to a constant)
Case 1: One of the variable is either absent or cancells out from the set of auxiliary
equations.
Case 2: if u=c1 is known but v=c2 is not possible by case 1, then use u=c1 to get v=c2
Case 3: Introducing Lagrange multipliers P1, Q1, R1 which are functions of x,y z or
constants, each fraction in (3) is equal to:

𝑃1 𝑑𝑥+𝑄1 𝑑𝑦+𝑅1 𝑑𝑧
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)
𝑃1 𝑃+𝑄1 𝑄+𝑄1 𝑅

If P1 , Q1 , R1 are chosen that P1P + Q 1Q + R1R=0 then P1dx + Q1dy + R1dz = 0

which can be integrated.
Case 4: Multipliers are chosen (more than once) such that the numerator P 1dx +
Q1dy + R1dz is an exact differential of denominator P1P + Q 1Q + R1R. Now combine
(4) with a fraction of (3) to get an integral.
4) General solution of (1) is F(u,v) = 0 or v = φ(u).

Examples:
𝜕𝑧 𝜕𝑧
1) Solve: (𝑚𝑧 − 𝑛𝑦)
𝜕𝑥
+ (𝑛𝑥 − 𝑙𝑧)
𝜕𝑦
= 𝑙𝑦 − 𝑚𝑥

Solution: The given equation is of the form Pp+Qq=R

The auxiliary equations are
𝑑𝑥 𝑑𝑦 𝑑𝑧
(𝑚𝑧−𝑛𝑦)
= (𝑛𝑥−𝑙𝑧) = (𝑙𝑦−𝑚𝑥) --------------- (1)

Using multipliers l, m, n each ratio is equal to

 𝑙𝑑𝑥+𝑚𝑑𝑦+𝑛𝑑𝑧 𝑙𝑑𝑥+𝑚𝑑𝑦+𝑛𝑑𝑧
=
𝑙(𝑚𝑧−𝑛𝑦)+𝑚(𝑛𝑥−𝑙𝑧)+𝑛(𝑙𝑦−𝑚𝑥) 0
⇒ 𝑙𝑑𝑥 + 𝑚𝑑𝑦 + 𝑛𝑑𝑧 = 0
which on integration gives 𝑙𝑥 + 𝑚𝑦 + 𝑛𝑧 = 𝑐1
𝑥𝑑𝑥+𝑦𝑑𝑦+𝑧𝑑𝑧
Using multipliers x, y, z each ratio in (1) is equal to =
𝑥(𝑚𝑧−𝑛𝑦)+𝑦(𝑛𝑥−𝑙𝑧)+𝑧(𝑙𝑦−𝑚𝑥)
𝑥𝑑𝑥+𝑦𝑑𝑦+𝑧𝑑𝑧
0
⇒ 𝑥𝑑𝑥 + 𝑦𝑑𝑦 + 𝑧𝑑𝑧 = 0
which on integration gives
𝑥2 𝑦2 𝑧2
+ + = 𝑐2 𝑜𝑟𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑐2
2 2 2
Hence the general solution is φ (lx + my + nz , x2 + y2 + z2)=0

2) Solve: (𝑥 2 − 𝑦𝑧)𝑝 + (𝑦 2 − 𝑧𝑥)𝑞 = 𝑧 2 − 𝑥𝑦

Solution: The given equation is of the form Pp+Qq=R
The auxiliary equations are
𝑑𝑥 𝑑𝑦 𝑑𝑧
(𝑥 2 −𝑦𝑧)
= (𝑦2 =
−𝑧𝑥) (𝑧 2 −𝑥𝑦)

The subsidiary equation is given by

𝑑𝑥 − 𝑑𝑦 𝑑𝑦 − 𝑑𝑧
=
(𝑥 2 − 𝑦𝑧) − (𝑦 2 − 𝑧𝑥) (𝑦 2 − 𝑧𝑥) − (𝑧 2 − 𝑥𝑦)
𝑑𝑧 − 𝑑𝑥
= 2 . . . . . . . . . . . . . . . (1)
(𝑧 − 𝑥𝑦) − (𝑥 2 − 𝑦𝑧)
Consider first two equations of (1) we get
𝑑𝑥 − 𝑑𝑦 𝑑𝑦 − 𝑑𝑧
⇒ =
(𝑥 2 − 𝑦𝑧) − (𝑦 2 − 𝑧𝑥) (𝑦 2 − 𝑧𝑥) − (𝑧 2 − 𝑥𝑦)

𝑑𝑥 − 𝑑𝑦 𝑑𝑦 − 𝑑𝑧
⇒ =
(𝑥 − 𝑦)(𝑥 + 𝑦 + 𝑧) (𝑦 − 𝑧)(𝑥 + 𝑦 + 𝑧)
𝑑𝑥 − 𝑑𝑦 𝑑𝑦 − 𝑑𝑧
⇒ =
(𝑥 − 𝑦) (𝑦 − 𝑧)
 𝑥−𝑦
⇒ 𝑙 𝑜𝑔( 𝑥 − 𝑦) = 𝑙𝑜𝑔(𝑦 − 𝑧) + 𝑙𝑜𝑔 𝑐1 ⇒ = 𝑐1
𝑦−𝑧
Similarly considering last two equations of (1) we get
⇒ 𝑙 𝑜𝑔( 𝑦 − 𝑧) = 𝑙𝑜𝑔(𝑧 − 𝑥) + 𝑙𝑜𝑔 𝑐2
𝑦−𝑧
⇒ = 𝑐2
𝑧−𝑥
Hence the required solution is
𝑥−𝑦 𝑦−𝑧
𝜑( , )=0
𝑦−𝑧 𝑧−𝑥

Method of Separation of Variables:

Separation of variables is a powerful technique to solve PDE. For a PDE in the
function u of two independent variables x and y, assume that the required solution
is separable i.e.
u (x,y)=X(x)Y(y) ---------(1)
where X(x) is a function of x alone and Y(y) is a function of y alone. Then
substitution of u from (1) and its derivatives reduces the PDE to the form;
f (X, X’, X’’,…...) = g(Y,Y’,Y’’,…..) ---------- (2)
which is separable in X and Y. Since the LHS of (2) is a function of x alone and RHS
of (2) is a function of y alone, then (2) must be equal to a common constant say “k”.
Thus (2) reduces to;
f (X,X’,X’’…...)=k ----------(3)
g (X,X’,X’’…...)=k ----------(4)
Thus the determination of solution to PDE reduces to the determination of
solutions to two ODE (with appropriate conditions).
Examples:
1. By the method of separation of variables, solve the equation:
𝜕𝑧 𝜕𝑧
𝑦3 + 𝑥2 = 0
𝜕𝑥 𝜕𝑦

Solution:
𝜕𝑧 3 𝜕𝑧 2
𝑦 + 𝑥 = 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
𝜕𝑥 𝜕𝑦
Let the solution of (1) be in the form z=XY where X=X(x) and Y=Y(y)
𝜕𝑧 𝜕𝑋 𝜕𝑧 𝜕𝑌
=𝑌 , =𝑋
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕𝑋 𝜕𝑌
(1) ⇒ ( 𝑌) 𝑦 3 + (𝑋 ) 𝑥 2 = 0
𝜕𝑥 𝜕𝑦
𝑑𝑋 𝑑𝑌
( 𝑌) 𝑦 3 = − (𝑋 ) 𝑥 2
𝑑𝑥 𝑑𝑦
1 𝑑𝑋 1 1 𝑑𝑌 1
( ) 2=− ( ) 3
𝑋 𝑑𝑥 𝑥 𝑌 𝑑𝑦 𝑦
L.H.S is a function of x only and R.H.S is a function of y only.
Since x and y are independent variables, this expression can hold only if each side
is a constant i.e.
1 𝑑𝑋 1 1 𝑑𝑌 1
( ) = 𝑘𝑎𝑛𝑑 − ( ) =𝑘
𝑋 𝑑𝑥 𝑥2 𝑌 𝑑𝑦 𝑦 3

where k is a constant. These may be written as

1 𝑑𝑋 1 𝑑𝑌
( ) = 𝑘𝑥 2 𝑎𝑛𝑑 ( ) = −𝑘𝑦 3
𝑋 𝑑𝑥 𝑌 𝑑𝑦
𝑑 𝑑
(𝑙𝑜𝑔 𝑋) = 𝑘𝑥 2 𝑎𝑛𝑑 (𝑙𝑜𝑔 𝑌) = −𝑘𝑦 3
𝑑𝑥 𝑑𝑦

Integrating w.r.t x and y we get,

 𝑥3 𝑦4
𝑙𝑜𝑔 𝑋 = 𝑘 + 𝑙𝑜𝑔 𝐶1 𝑎𝑛𝑑 𝑙𝑜𝑔 𝑌 = −𝑘 + 𝑙𝑜𝑔 𝐶2
3 4
𝑘𝑥 3 −𝑘𝑦 4
⇒𝑋= 𝐶1 𝑒 3 𝑎𝑛𝑑𝑌 = 𝐶2 𝑒 4 𝑤ℎ𝑒𝑟𝑒𝐶1 𝑎𝑛𝑑𝐶2 𝑎𝑟𝑒𝑐𝑜𝑛𝑠 𝑡𝑎𝑛 𝑡 𝑠
𝑥3 𝑦4
𝑘3 −𝑘
∴ 𝑧 = 𝑋𝑌 = 𝐴𝑒 𝑒 4 𝑤ℎ𝑒𝑟𝑒𝐴 = 𝐶1 𝐶2

2. Using method of separation of variables, solve

𝜕𝑢 𝜕𝑢
=2 + 𝑢𝑤ℎ𝑒𝑟𝑒𝑢 (𝑥, 0) = 6𝑒 −3𝑥
𝜕𝑥 𝜕𝑡
Solution: Here x and t are independent variables and u is the dependent variable.
Let the solution be in the form u=X(x)Y(t)
𝜕𝑢 𝑑𝑋 𝜕𝑢 𝑑𝑋
∴ = 𝑌(𝑡)𝑎𝑛𝑑 = 𝑋(𝑥 )
𝜕𝑥 𝑑𝑥 𝜕𝑡 𝑑𝑡

The equation takes the form

𝑑𝑋 𝑑𝑌
𝑌(𝑡) = 2𝑋(𝑥 ) + 𝑋(𝑥 )𝑌(𝑦)
𝑑𝑥 𝑑𝑡
𝑑𝑋 𝑑𝑌 1 𝑑𝑋 2 𝑑𝑌
( − 𝑋) 𝑌 = 2𝑋 ⇒ −1=
𝑑𝑥 𝑑𝑡 𝑋 𝑑𝑥 𝑌 𝑑𝑡
1 𝑑𝑋 2 𝑑𝑌 𝑑𝑋 𝑑𝑌 𝑘
− 1 = 𝑘𝑎𝑛𝑑 =𝑘⇒ = (𝑘 + 1)𝑑𝑥𝑎𝑛𝑑 = 𝑑𝑡
𝑋 𝑑𝑥 𝑌 𝑑𝑡 𝑋 𝑌 2

On integration we get,
𝑘
𝑙𝑜𝑔 𝑋 = (𝑘 + 1)𝑥 + 𝑙𝑜𝑔 𝐶1 𝑎𝑛𝑑 𝑙𝑜𝑔 𝑌 = 𝑡 + 𝑙𝑜𝑔 𝐶2
2
𝑘
(𝑘+1)𝑥 𝑡
⇒ 𝑋 = 𝐶1 𝑒 𝑎𝑛𝑑𝑋 = 𝐶2 𝑒 2

𝑘
Therefore the required solution is𝑢 = 𝑋𝑌 = 𝐴𝑒 (𝑘+1)𝑥 𝑒 2𝑡 𝑤ℎ𝑒𝑟𝑒𝐴 =
𝐶1 𝐶2
But given u(x,0)=6e-3x
=> 6e-3x =Ae (k+1)x
=> A=6 and k+1 = -3
=> A=6 and k=-4
∴ 𝑢 = 6𝑒 −(3𝑥+2𝑡)

CLASSIFICATION OF PARTIAL DIFFERENTIAL EQUATIONS OF

SECOND ORDER
The general form of a second-order PDE in the function u of the two independent
variables x, y is given by

𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢 𝜕𝑢 𝜕𝑢
𝐴(𝑥, 𝑦) 2
+ 𝐵(𝑥, 𝑦) + 𝐶(𝑥, 𝑦) + 𝑓 (𝑥, 𝑦, 𝑢, , )=0 (1)
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦 2 𝜕𝑥 𝜕𝑦

This equation is linear in second order terms. PDE (1) is said to be “linear or quasi-
linear” according as f is linear or non-linear.
PDE (1) is classified as Elliptic, Parabolic or Hyperbolic according to B2 -4AC <
0, =0 or > 0
Example:
Elliptic: (B2 -4AC < 0)

Laplace’s equation in two dimensions

𝜕2 𝑢 𝜕2 𝑢
2
+ = 0 ………………………….. (2)
𝜕𝑥 𝜕𝑦 2

Poisson’s equation in two dimensions

𝜕2 𝑢 𝜕2 𝑢
2
+ = 𝑓(𝑥, 𝑦) … … … … … … … … … … . . (3)
𝜕𝑥 𝜕𝑦 2

Parabolic: (B2 -4AC = 0)

One dimensional heat flow equation
𝜕2 𝑢 𝜕𝑢
𝑎2 = ………………………….. (4)
𝜕𝑥 2 𝜕𝑡

Hyperbolic: (B2 -4AC > 0)

One dimensional wave equation
 𝐿
2 𝑛πx
where An = ∫ 𝑓(𝑥). sin 𝑑𝑥
𝐿 𝐿
0

2 𝑏𝑑 𝑛πx 2 𝑑 𝐿
𝑛πx
𝐴𝑛 = ∫ 𝑥. sin ( ) 𝑑𝑥 + ∫ (𝑥 − 𝐿)sin 𝑑𝑥
𝐿 0 𝑏 𝐿 𝐿 (𝑏 − 𝐿) 𝑏 𝐿
𝑏
2𝑑 𝐿 𝑛πx 𝐿2 𝑛πx 2𝑑 𝐿 𝑛πx
= [𝑥. (− ) . cos − (− ) . sin ] + [(𝑥 − 𝐿) (− ) cos −
𝑏𝐿 𝑛π 𝐿 𝑛2 π2 𝐿 0 𝐿(𝑏−𝐿) 𝑛π 𝐿
𝐿
𝐿2 𝑛πx
(− 𝑛2π2 ) . sin 𝐿
]
𝑏
2𝑑𝑏𝐿 𝑛πb 2𝑑𝐿2 𝑛πb 2𝑑 𝑛πb 2𝑑𝐿2 𝑛πb
=− . cos + . sin + . cos − . sin
𝑏𝐿𝑛π 𝐿 𝑏𝐿𝑛2 π 2 𝐿 𝑛π 𝐿 𝐿(𝑏−𝐿)𝑛2 π2 𝐿

2𝑑𝐿2 𝑛πb
Thus An = . sin
𝑏(𝐿−𝑏)𝑛2 π 2 𝐿

Hence the subsequent displacement of the string is given by the displacement

function
∞
2𝑑𝐿2 1 𝑛πb 𝑛𝜋𝑥 𝑛𝜋𝑎𝑡
𝑦(𝑥, 𝑡) = ∑ sin . sin . cos
𝑏(𝐿 − 𝑏)π2 𝑛2 𝐿 𝐿 𝐿
𝑛=1

Homogeneous linear partial differential equations with constant

coefficients

Consider a partial differential equation of the form

𝜕𝑛𝑧 𝜕𝑛𝑧 𝜕𝑛𝑧 𝜕𝑛 𝑧
𝐴0 + 𝐴1 + 𝐴 2 +. . . . . +𝐴𝑛 +⋯
𝜕𝑥 𝑛 𝜕𝑥 𝑛−1 𝜕𝑦 𝜕𝑥 𝑛−2 𝜕𝑦 2 𝜕𝑦 𝑛
= 𝐹(𝑥, 𝑦). . . . . . . . . . . (1)
where A0, A1 …., An are constant coefficients. In this equation the dependent variable
z and its derivatives are linear. Since each term (in the LHS) of (1) contains z or its
derivatives, equation (1) is known as homogeneous linear partial differential
equation of order n with constant coefficients.
𝜕 𝜕
For convenience and will be denoted by D or Dx and D1 or Dy respectively.
𝜕𝑥 𝜕𝑦
Then (1) can be rewritten as:
(𝐴0 𝐷𝑥 𝑛 + 𝐴1 𝐷𝑥 𝑛−1 𝐷𝑦 + ⋯ . . +𝐴𝑛 𝐷𝑦 𝑛 )𝑧 = 𝐹(𝑥, 𝑦)
The above equation may rewritten as:
𝐹(𝐷𝑥 , 𝐷𝑦 )𝑧 = 𝐹(𝑥, 𝑦). . . . . . . . . (2)

Where 𝐹(𝐷𝑥 , 𝐷𝑦 ) = 𝐴0 𝐷𝑥 𝑛 + 𝐴1 𝐷𝑥 𝑛−1 𝐷𝑦 +. . . . . +𝐴𝑛 𝐷𝑦 𝑛

Remark:
Equation (1) is called as homogeneous because all terms contain derivatives of same
order.

Alternative definition
𝜕𝑛 𝑧 𝜕𝑛𝑧 𝜕𝑛 𝑧 𝜕𝑛 𝑧
+ 𝑘1 𝑛−1 + 𝑘2 𝑛−2 2 +. . . . . +𝑘𝑛 𝑛 + ⋯ = 𝐹(𝑥, 𝑦). . . . . . . . . . . (1)
𝜕𝑥 𝑛 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑦

where k’s are constant, is called a homogeneous linear partial differential equation
of the nth order with constant coefficients. Equation (1) is called as homogeneous
because all terms contain derivatives of same order.

𝜕𝑟 𝜕𝑟
On writing, = 𝐷𝑟 and = 𝐷′𝑟 .
𝜕𝑥 𝑟 𝜕𝑦 𝑟

Equation (1) becomes (𝐷𝑛 + 𝑘1 𝐷𝑛−1 𝐷′ + ⋯ … … … + 𝑘𝑛 𝐷′𝑛 )𝑍 = 𝐹(𝑥, 𝑦)

Or briefly
𝑓(D , 𝐷′ )𝑧 = 𝐹(𝑥, 𝑦). . . . . . . . . (2)

As in case of ordinary linear equations with constant coefficients the complete

solution of (1) consists of two parts, namely the complementary function and the
particular integral.
The complementary function is the complete solution of the equation 𝑓(D , 𝐷′ )𝑧 =
0, which must contain n arbitrary functions. The particular integral is the particular
solution of equation (2).
For example:
𝜕2 𝑧 𝜕2 𝑧 𝜕2 𝑧
1. 3 2
+5 + =0 Homogeneous PDE of order 2.
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦 2
𝜕3 𝑧 𝜕3 𝑧 𝜕3 𝑧 𝜕3 𝑧
2. 2 3
−3 +5 −8 = 𝑒 𝑥+𝑦 Homogeneous PDE of order 3.
𝜕𝑥 𝜕𝑥 2 𝜕𝑦 𝜕𝑥𝜕𝑦 2 𝜕𝑦 3
𝜕2 𝑧 𝜕2 𝑧 𝜕2 𝑧
3. 2
−5 +3 = 𝑥 + 𝑦 Homogeneous PDE of order 2.
𝜕𝑥 𝜕𝑥𝜕𝑦 𝜕𝑦 2

Complementary functions (C.F.) of homogeneous linear partial differential

equation i.e. 𝐹(𝐷𝑥 , 𝐷𝑦 )𝑧 = 0
Let 𝑧 = ∅(𝑦 + 𝑚𝑥) be a solution (2) where ∅ is an arbitrary function and m is is a
constant, then
𝐷𝑥 𝑧 = 𝜙′(𝑦 + 𝑚𝑥) · 𝑚
𝐷𝑥 2 𝑧 = 𝜙′′(𝑦 + 𝑚𝑥) · 𝑚2
.
.
.
𝐷𝑥 𝑛 𝑧 = 𝜙 (𝑛) (𝑦 + 𝑚𝑥) · 𝑚𝑛

𝐷𝑦 𝑧 = 𝜙′(𝑦 + 𝑚𝑥)

𝐷𝑦 2 𝑧 = 𝜙′′(𝑦 + 𝑚𝑥)
.
.
𝐷𝑦 𝑛 𝑧 = 𝜙 (𝑛) (𝑦 + 𝑚𝑥)
𝐷𝑥 𝐷𝑦 𝑧 = 𝑚 𝜙′′(𝑦 + 𝑚𝑥)

𝐷2 𝑥 𝐷𝑦 𝑧 = 𝑚2 𝜙 (3) (𝑦 + 𝑚𝑥)
.
 .
.
𝐷𝑟 𝑥 𝐷 𝑠 𝑦 𝑧 = 𝑚𝑟 𝜙 (𝑟+𝑠) (𝑦 + 𝑚𝑥)

= 𝑚𝑟 𝜙 (𝑛) (𝑦 + 𝑚𝑥)
where r + s = n
Substituting these values in (2) and simplifying, we get:
(𝐴0 𝑚𝑛 + 𝐴1 𝑚𝑛−1 +. . . . . +𝐴𝑛 )𝜙 (𝑛) (𝑦 + 𝑚𝑥) = 0. . . . . . . (4)
Which is true if m is a root of the equation:
𝐴0 𝑚𝑛 + 𝐴1 𝑚𝑛−1 +. . . . . +𝐴𝑛 = 0. . . . . . . (5)
The equation (5) is known as the (characteristic equation) or the (Auxiliary equation
(A.E.)) and is obtained by putting Dx = m and Dy=1
In 𝐹(𝐷𝑥 , 𝐷𝑦 )𝑧 = 0, and it has n roots.
Let m0, m1 …., mn be n roots of A.E. (5).
Three cases arise:
Case 1: When the roots are distinct:
If m0, m1 …., mn be n distinct roots of A.E. (5) then
𝜙1 (𝑦 + 𝑚1 𝑥), 𝜙2 (𝑦 + 𝑚2 𝑥), . . . . . . . . . , 𝜙𝑛 (𝑦 + 𝑚𝑛 𝑥) are the linear solution
corresponding to them and since the sum of any linear solutions is a solution too
than the general solution in this case is:
𝑧 = 𝜙1 (𝑦 + 𝑚1 𝑥) + 𝜙2 (𝑦 + 𝑚2 𝑥)+ . . . . . . . . . + 𝜙𝑛 (𝑦 + 𝑚𝑛 𝑥) . . . . . . (6)

Example:
Find the general solution of:
(𝐷𝑥 3 + 2𝐷𝑥 2 𝐷𝑦 − 5𝐷𝑥 𝐷𝑦 2 − 6𝐷𝑦 3 )𝑧 = 0
Solution. The A.E. is 𝑚3 + 2𝑚2 − 5𝑚 − 6 = 0
→ (𝑚 + 1)(𝑚2 + 𝑚 − 6) = 0
 → (𝑚 + 1)(𝑚 + 3)(𝑚 − 2) = 0
𝑚1 = −1, 𝑚2 = −3, 𝑚3 = 2
Note that 𝑚1 , 𝑚2 and m3 are different roots, then the general solution is
𝑧 = 𝜙1 (𝑦 + 𝑚1 𝑥) + 𝜙2 (𝑦 + 𝑚2 𝑥) + 𝜙3 (𝑦 + 𝑚3 𝑥)
→ 𝑧 = 𝜙1 (𝑦 − 𝑥) + 𝜙2 (𝑦 − 3𝑥) + 𝜙3 (𝑦 + 2𝑥)
Where 𝜙1 , 𝜙2 𝑎𝑛𝑑 𝜙3 are arbitrary functions.

Case 2: When the roots are repeated.

If the root m is repeated k times i.e. 𝑚1 = 𝑚2 =. . . = 𝑚𝑘 , then the
corresponding solution is :
𝑧 = 𝜙1 (𝑦 + 𝑚1 𝑥) + 𝑥𝜙2 (𝑦 + 𝑚1 𝑥)+. . . . +𝑥 𝑘−1 𝜙𝑛 (𝑦 + 𝑚1 𝑥). . . . . (7)
where 𝜙1 , . . . , 𝜙𝑘 are arbitrary functions.
Note: If some of the roots m0, m1 …., mn are repeated and the other are not i.e.
𝑚1 = 𝑚2 =. . . = 𝑚𝑘 ≠ 𝑚𝑘+1 ≠. . . . ≠ 𝑚𝑛 then the general solution is :
𝑧 = 𝜙1 (𝑦 + 𝑚1 𝑥) + 𝑥𝜙2 (𝑦 + 𝑚1 𝑥)+. . . + 𝑥 𝑘−1 𝜙𝑘 (𝑦 + 𝑚1 𝑥)
+ 𝜙𝑘+1 (𝑦 + 𝑚𝑘+1 𝑥)+. . . +𝜙𝑛 (𝑦 + 𝑚𝑛 𝑥) . . . . . . (8)
Example:
Solve (𝐷𝑥 3 − 𝐷𝑥 2 𝐷𝑦 − 8𝐷𝑥 𝐷𝑦 2 + 12𝐷𝑦 3 )𝑧 = 0
Solution: The A.E. is 𝑚3 − 𝑚2 − 8𝑚 + 12 = 0
→ (𝑚 − 2)(𝑚 − 2)(𝑚 + 3) = 0
m1 = m2 = 2, m3 = -3
Then, the general solution is
𝑧 = 𝜙1 (𝑦 + 2𝑥) + 𝑥𝜙2 (𝑦 + 2𝑥) + 𝜙3 (𝑦 − 3𝑥)
Where 𝜙1 , 𝜙2 and 𝜙3 are arbitrary functions.
Case 3: When the roots are complex.
If one of the roots of the given equation is complex let be m1 , then the
conjugate of m1 is also a root, let be m2 , so the general solution is:
𝑧 = 𝜙1 (𝑦 + 𝑚1 𝑥) + 𝜙2 (𝑦 + 𝑚2 𝑥)+ . . . . . . . . . + 𝜙𝑛 (𝑦 + 𝑚𝑛 𝑥)
Where 𝜙1 , . . . , 𝜙𝑛 are arbitrary functions.

Example:
Solve (𝐷𝑥 2 − 2𝐷𝑥 𝐷𝑦 + 5𝐷𝑦 2 )𝑧 = 0
Solution: The A. E. is 𝑚2 − 3𝑚 + 5 = 0
2 ± √4 − 20
→𝑚= = 1 ± 2𝑖
2
∴ 𝑚1 = 1 + 2𝑖 , 𝑚2 = 1 − 2𝑖
𝑧 = 𝜙1 (𝑦 + (1 + 2𝑖)𝑥) + 𝜙2 (𝑦 + (1 − 2𝑖)𝑥)
That is the general solution where 𝜙1 , 𝜙2 are arbitrary functions.

Particular integral (P.I.) of partial differential equation:

When 𝑓(𝑥, 𝑦) ≠ 0 in the equation (2) which it's 𝑓(𝐷𝑥 , 𝐷𝑦 )𝑧 = 𝐹(𝑥, 𝑦).
1
Multiplying (3) by the inverse operator 𝑓(𝐷 ,𝐷 )
𝑥 𝑦

1 1
· 𝑓(𝐷𝑥 , 𝐷𝑦 )𝑧 = 𝐹(𝑥, 𝑦)
𝑓(𝐷𝑥 , 𝐷𝑦 ) 𝑓(𝐷𝑥 , 𝐷𝑦 )
1
𝑧= 𝐹(𝑥, 𝑦) …….. (4)
𝑓(𝐷𝑥 ,𝐷𝑦 )

which is the particular integral (P.I.)

The operator 𝐹(𝐷𝑥 , 𝐷𝑦 ) can be written as
𝑓(𝐷𝑥 , 𝐷𝑦 ) = (𝐷𝑥 − 𝑚1 𝐷𝑦 )(𝐷𝑥 − 𝑚2 𝐷𝑦 ). . . (𝐷𝑥 − 𝑚𝑛 𝐷𝑦 ) . . . . . (5)
Substituting (5) in (4) :
1
𝑧= 𝐹(𝑥, 𝑦)
(𝐷𝑥 − 𝑚1 𝐷𝑦 )(𝐷𝑥 − 𝑚2 𝐷𝑦 ). . . (𝐷𝑥 − 𝑚𝑛 𝐷𝑦 )
1
Taking 𝑢1 = (𝐷 𝐹(𝑥, 𝑦)
𝑥 −𝑚𝑛 𝐷𝑦 )

∴ (𝐷𝑥 − 𝑚𝑛 𝐷𝑦 )𝑢1 = 𝐹(𝑥, 𝑦)

This equation can be solved by Lagrange's method.

The Lagrange's auxiliary equations are:
𝑑𝑥 𝑑𝑦 𝑑𝑢1
= = . . . . . . . . (6)
1 −𝑚𝑛 𝐹(𝑥, 𝑦)
Taking the first two fractions of (6)
𝑚𝑛 𝑑𝑥 + 𝑑𝑦 = 0 → 𝑚𝑛 𝑥 + 𝑦 = 𝑎
Taking the first and third fractions of (6)
𝑑𝑢1
𝑑𝑥 = → 𝐹(𝑥, 𝑦)𝑑𝑥 = 𝑑𝑢1
𝐹(𝑥, 𝑦)
On substitution we get
𝐹(𝑥, 𝑎 − 𝑚𝑛 𝑥)𝑑𝑥 = 𝑑𝑢1
Integrating the above equation we get
𝑢1 = ∫ 𝐹(𝑥, 𝑎 − 𝑚𝑛 𝑥)𝑑𝑥 + 𝑏
Let b=0, then we have u1
Similarly, we take
1
𝑢2 = 𝑢
𝐷𝑥 − 𝑚𝑛−1 𝐷𝑦 1

And solve it by Lagrange's method, we get u2,

1
𝑧 = 𝑢2 = 𝑢
𝐷𝑥 − 𝑚1 𝐷𝑦 𝑛−1
And by solving this equation we get the particular integral (P.I.).

Short methods of finding the P.I. in certain cases:

Case 1: When 𝐹(𝑥, 𝑦) = 𝑒 𝑎𝑥+𝑏𝑦 where a and b are arbitrary constants:
To find the P.I. when 𝐹(𝑎, 𝑏) ≠ 0, we derive 𝐹(𝑥, 𝑦) for x and y n times:
𝐷𝑥 𝑒 𝑎𝑥+𝑏𝑦 = 𝑎𝑒 𝑎𝑥+𝑏𝑦
𝐷𝑥 2 𝑒 𝑎𝑥+𝑏𝑦 = 𝑎2 𝑒 𝑎𝑥+𝑏𝑦
.
.
𝐷𝑥 𝑛 𝑒 𝑎𝑥+𝑏𝑦 = 𝑎𝑛 𝑒 𝑎𝑥+𝑏𝑦

𝐷𝑦 𝑒 𝑎𝑥+𝑏𝑦 = 𝑏𝑒 𝑎𝑥+𝑏𝑦
𝐷2 𝑦 𝑒 𝑎𝑥+𝑏𝑦 = 𝑏 2 𝑒 𝑎𝑥+𝑏𝑦
.
.
𝐷𝑛 𝑦 𝑒 𝑎𝑥+𝑏𝑦 = 𝑏 𝑛 𝑒 𝑎𝑥+𝑏𝑦

𝐷𝑥 𝑟 𝐷𝑦 𝑠 𝑒 𝑎𝑥+𝑏𝑦 = 𝑎𝑟 𝑏 𝑠 𝑒 𝑎𝑥+𝑏𝑦 where 𝑟 + 𝑠 = 𝑛

So
𝑓(𝐷𝑥 , 𝐷𝑦 )𝑒 𝑎𝑥+𝑏𝑦 = 𝑓(𝑎, 𝑏)𝑒 𝑎𝑥+𝑏𝑦
1
Multiplying both sides by , we get
𝑓(𝐷𝑥 ,𝐷𝑦 )

1
𝑒 𝑎𝑥+𝑏𝑦 = 𝑓(𝑎, 𝑏)𝑒 𝑎𝑥+𝑏𝑦
𝑓(𝐷𝑥 , 𝐷𝑦 )

Since 𝑓(𝑎, 𝑏) ≠ 0 then we can divide on it:

 1 1
𝑒 𝑎𝑥+𝑏𝑦 = 𝑒 𝑎𝑥+𝑏𝑦 ……… (*)
𝑓(𝑎,𝑏) 𝑓(𝐷𝑥 ,𝐷𝑦 )

Which it is equal to 𝑧, then the P. I. is

1 1
𝑧= 𝑒 𝑎𝑥+𝑏𝑦 = 𝑒 𝑎𝑥+𝑏𝑦 , where 𝑓(𝑎, 𝑏) ≠ 0
𝑓(𝐷𝑥 ,𝐷𝑦 ) 𝑓(𝑎.𝑏)

When (𝑎, 𝑏) = 0 , then analyze 𝑓(𝐷𝑥 , 𝐷𝑦 ) as follows

𝑎 𝑟
𝑓(𝐷𝑥 , 𝐷𝑦 ) = (𝐷𝑥 − 𝐷𝑦 ) 𝐺(𝐷𝑥 , 𝐷𝑦 )
𝑏
where G (a, b) ≠ 0 we get
1 1
𝑧= 𝑒 𝑎𝑥+𝑏𝑦 = 𝑟 𝑒 𝑎𝑥+𝑏𝑦
𝑓(𝐷𝑥 , 𝐷𝑦 ) 𝑎
(𝐷𝑥 − 𝐷𝑦 ) 𝐺(𝐷𝑥 , 𝐷𝑦 )
𝑏
1 1
= 𝑎 𝑟 · 𝑒 𝑎𝑥+𝑏𝑦 from (*)
(𝐷𝑥 − 𝐷𝑦 ) 𝐺(𝑎,𝑏)
𝑏

Since G (a, b) ≠ 0
1 1 𝑎𝑥+𝑏𝑦
= · 𝑟𝑒
𝑎
𝐺(𝐷𝑥 , 𝐷𝑦 ) (𝐷 − 𝐷 )
𝑥 𝑏 𝑦
Then by Lagrange's method r times , we get
1 𝑎𝑥+𝑏𝑦
1 𝑥 𝑟 𝑎𝑥+𝑏𝑦
𝑧= 𝑒 = · 𝑒
𝑓(𝐷𝑥 , 𝐷𝑦 ) 𝐺(𝑎, 𝑏) 𝑟!

Which it's the P.I. where

𝑓(𝑎, 𝑏) = 0, 𝐺(𝑎, 𝑏) ≠ 0

Example: Solve (𝑫𝒙 𝟐 − 𝑫𝒙 𝑫𝒚 − 𝟔𝑫𝒚 𝟐 )𝒛 = 𝒆𝟐𝒙−𝟑𝒚

Solution.
1) To find the general solution
The A.E. of the given equation is
 𝑚2 − 𝑚 − 6 = 0 → (𝑚 − 3)(𝑚 + 2) = 0
∴ 𝑚1 = 3, 𝑚2 = −2
∴ 𝑧1 = 𝜙1 (𝑦 + 3𝑥) + 𝜙2 (𝑦 − 2𝑥)
where 𝜙1 and 𝜙2 are arbitrary functions
2) To find the particular Integral (P.I.)
a = 2, b = -3
𝐹(𝑎, 𝑏) = 𝑎2 − 𝑎𝑏 − 6𝑏 2
𝐹(2, −3) = 4 + 6 − 54 = −44 ≠ 0
1 1 2𝑥−3𝑦
𝑧2 = 𝑒 𝑎𝑥+𝑏𝑦 = 𝑒
𝐹(𝑎. 𝑏) −44
∴ The general solution is z=C.F+P.I
1
i.e. 𝑧 = 𝑧1 + 𝑧2 = 𝜙1 (𝑦 + 3𝑥) + 𝜙2 (𝑦 − 2𝑥) − 𝑒 2𝑥−3𝑦
44

Case 2: When 𝐹(𝑥, 𝑦) = sin(𝑎𝑥 + 𝑏𝑦) 𝑜𝑟 cos(𝑎𝑥 + 𝑏𝑦) where a and b are
arbitrary constants
Let 𝐹(𝑥, 𝑦) = sin(𝑎𝑥 + 𝑏𝑦)
𝐷𝑥 sin(𝑎𝑥 + 𝑏𝑦) = 𝑎cos(𝑎𝑥 + 𝑏𝑦)
𝐷𝑥 2 𝑠𝑖𝑛(𝑎𝑥 + 𝑏𝑦) = −𝑎2 𝑠𝑖𝑛(𝑎𝑥 + 𝑏𝑦)
𝐷𝑦 𝑠𝑖𝑛(𝑎𝑥 + 𝑏𝑦) = 𝑏𝑐𝑜𝑠(𝑎𝑥 + 𝑏𝑦)

𝐷𝑦 2 sin(𝑎𝑥 + 𝑏𝑦) = −𝑏 2 sin(𝑎𝑥 + 𝑏𝑦)

𝐷𝑥 𝐷𝑦 sin(𝑎𝑥 + 𝑏𝑦) = 𝐷𝑥 [𝑏cos(𝑎𝑥 + 𝑏𝑦)]
= −𝑎𝑏sin(𝑎𝑥 + 𝑏𝑦)
𝑓(𝐷𝑥 2 , 𝐷𝑥 𝐷𝑦 , 𝐷𝑦 2 )sin(𝑎𝑥 + 𝑏𝑦) = 𝑓(−𝑎2 , −𝑎𝑏, −𝑏 2 )sin(𝑎𝑥 + 𝑏𝑦)
 1
Multiplying both sides by 𝐹(𝐷 2 2
𝑥 ,𝐷𝑥 𝐷𝑦 ,𝐷𝑦 )

If 𝑓(−𝑎2 , −𝑎𝑏, −𝑏 2 ) ≠ 0 then we can divide on it

1
→𝑧= sin(𝑎𝑥 + 𝑏𝑦)
𝑓(𝐷𝑥 2 , 𝐷𝑥 𝐷𝑦 , 𝐷𝑦 2 )
1
= sin(𝑎𝑥 + 𝑏𝑦)
𝑓(−𝑎2 , −𝑎𝑏, −𝑏 2 )
which it is the particular integral.
And if 𝑓(−𝑎2 , −𝑎𝑏, −𝑏2 ) = 0 then we write
𝑒 𝑖𝜃 − 𝑒 −𝑖𝜃 𝑒 𝑖𝜃 + 𝑒 −𝑖𝜃
sin 𝜃 = , cos 𝜃 =
2𝑖 2
And follow the solution of the exponential function in case 1.

Example: Solve (𝐷𝑥 2 − 3𝐷𝑥 𝐷𝑦 + 𝐷𝑦 2 )𝑧 = 𝑒 2𝑥+3𝑦 + 𝑒 𝑥+𝑦 + sin(𝑥 − 2𝑦)

Solution.
1) Finding the general solution z
The A.E. is
𝑚2 − 3𝑚 + 0 ⇒ (𝑚 − 2)(𝑚 − 1) = 0
∴ 𝑚1 = 2, 𝑚2 = 1
∴ 𝑧1 = 𝜙1 (𝑦 + 2𝑥) + 𝜙2 (𝑦 + 𝑥)
where 𝜙1 and 𝜙2 are arbitrary functions.
2) The P.I. of the given equation is
1 1 1
P.I. = 𝑧2 = 𝑒 2𝑥+3𝑦 + 𝑒 𝑥+𝑦 + sin(𝑥 − 2𝑦)
𝐹(𝐷𝑥 , 𝐷𝑦 ) 𝐹(𝐷𝑥 , 𝐷𝑦 ) 𝐹(𝐷𝑥 , 𝐷𝑦 )
1
Let 𝑢1 = 𝐹(𝐷 𝑒 2𝑥+3𝑦 , 𝑎 = 2, 𝑏 = 3
𝑥 ,𝐷𝑦 )
 𝐹(𝐷𝑥 , 𝐷𝑦 ) = 𝑎2 − 3𝑎𝑏 + 2𝑏 2
𝐹(1, 1) = 4 − 18 + 18 = 4 ≠ 0
1
𝑢1 = 𝑒 2𝑥+3𝑦
4
1
𝑢2 = 𝑒 𝑥+𝑦 , 𝑎 = 1, 𝑏 = 1
𝐹(𝐷𝑥 , 𝐷𝑦 )
𝐹(𝑎, 𝑏) = 𝑎2 − 3𝑎𝑏 + 2𝑏 2
𝐹(1, 1) = 1 − 3 + 2 = 0
Analyse 𝐹(𝐷𝑥 , 𝐷𝑦 ),
𝐹(𝐷𝑥 , 𝐷𝑦 ) = (𝐷𝑥 − 2𝐷𝑦 )(𝐷𝑥 − 𝐷𝑦 )
1 𝑥 𝑟 𝑎𝑥+𝑏𝑦
𝑢2 = 𝑒
𝐺(𝑎, 𝑏) 𝑟!
1 𝑥 𝑥+𝑦
= 𝑒
−1 1
𝑢2 = −𝑥𝑒 𝑥+𝑦
1
𝑢3 = sin(𝑥 − 2𝑦)
𝐹(𝐷𝑥 , 𝐷𝑦 )
𝐹(−𝑎2 , −𝑎𝑏, −𝑏 2 ) = −𝑎2 + 3𝑎𝑏 − 2𝑏 2
𝐹(−1, 2, −4) = −1 − 6 − 8 = −15 ≠ 0
1
𝑢3 = sin(𝑥 − 2𝑦)
−15
Then, the required general solution is
1 1
𝑧 = 𝑧1 + 𝑧2 = 𝜙1 (𝑦 + 2𝑥) + 𝜙2 (𝑦 + 𝑥) + 𝑒 2𝑥+3𝑦 − 𝑥𝑒 𝑥+𝑦 − sin(𝑥 − 2𝑦)
4 15
Where 𝜙1 and 𝜙2 are arbitrary functions.
Case 3: When 𝐹(𝑥, 𝑦) = 𝑥 𝑎 𝑦 𝑏 where a and b are Non- Negative Integer
Number
1
The particular integral (P.I.) is evaluated by expanding the function 𝑓(𝐷
𝑥 ,𝐷𝑦 )

in an infinite series of ascending powers of 𝐷𝑥 or 𝐷𝑦 (i.e) by transfer the

1
function 𝑓(𝐷 ,𝐷 ) according to the following
𝑥 𝑦

1
= 1 + 𝜃 + 𝜃 2 +. . . .
1−𝜃

Example: Find P.I. of the equation (𝐷𝑥 3 − 7𝐷𝑥 𝐷𝑦 2 − 6𝐷𝑦 3 )𝑧 = 𝑥 2 𝑦

1
Solution: P.I = 𝑥 2𝑦
𝐷𝑥 −7𝐷𝑥 𝐷𝑦 2 −6𝐷𝑦 3
3

1
= 2 3 𝑥 2𝑦
7𝐷𝑦 6𝐷𝑦
𝐷𝑥 3 [1 − ( 2 +
𝐷𝑥 𝐷𝑥 3 )]
2
1 7𝐷𝑦 2 6𝐷𝑦 3 7𝐷𝑦 2 6𝐷𝑦 3
= 3 [1 + ( 2 + )+( 2 + ) +. . . ] 𝑥 2 𝑦
𝐷𝑥 𝐷𝑥 𝐷 𝑥 3 𝐷𝑥 𝐷 𝑥 3

1
= [𝑥 2 𝑦]
𝐷𝑥 3
2
7𝐷𝑦 2 6𝐷𝑦 3 2 7𝐷𝑦 2 6𝐷𝑦 3
Since ( 𝐷 2 + 𝐷𝑥3
) (𝑥 𝑦) = 0, ( 𝐷 2 +
𝐷𝑥3
) (𝑥 2 𝑦) = 0
𝑥 𝑥

1 𝑥 3𝑦 1 𝑥 4𝑦 𝑥 5𝑦
= 2 = =
𝐷𝑥 3 𝐷𝑥 12 60

Case 4 When 𝐹(𝑥, 𝑦) = 𝑒 𝑎𝑥+𝑏𝑦 𝑉 where V is a function of x and y

1
The P.I. in this case is 𝑧 = 𝐹(𝐷 𝑒 𝑎𝑥+𝑏𝑦 𝑉
𝑥 ,𝐷𝑦 )

1
= 𝑒 𝑎𝑥+𝑏𝑦 𝑉
𝐹(𝐷𝑥 + 𝑎, 𝐷𝑦 + 𝑏)
and solving this equation depending on the type of V can get the
particular integral (P.I.), as follows:

Example: Find P.I. of the equation 𝐷𝑥 𝐷𝑦 𝑧 = 𝑒 2𝑥+3𝑦 𝑥 2 𝑦

1
Solution: P.I. = 𝐷 𝑒 2𝑥+3𝑦 𝑥 2 𝑦 𝑎 = 2, 𝑏 = 3 𝑎𝑛𝑑 𝑉 = 𝑥 2 𝑦
𝑥 𝐷𝑦

1
= 𝑒 2𝑥+3𝑦 𝑥 2𝑦
(𝐷𝑥 + 2)(𝐷𝑦 + 3)
1
= 𝑒 2𝑥+3𝑦 𝑥 2𝑦
𝐷𝑦
3(𝐷𝑥 + 2) (1 + )
3

2𝑥+3𝑦
1 𝐷𝑦 𝐷𝑦 2
=𝑒 [1 − + −. . . ] 𝑥 2 𝑦
3(𝐷𝑥 + 2) 3 9
1 𝑥2
= 𝑒 2𝑥+3𝑦 [𝑥 2
𝑦 − ]
3(𝐷𝑥 + 2) 3

2𝑥+3𝑦
1 2
𝑥2
=𝑒 [𝑥 𝑦 − ]
𝐷𝑥 3
6 (1 + )
2
1 2𝑥+3𝑦 𝐷𝑥 𝐷𝑥 2 𝐷𝑥 3 2
𝑥3 𝐷𝑥 3
= 𝑒 [1 − + − +. . . ] [𝑥 𝑦 − ] , ( = 0)
6 2 4 8 3 8
1 2𝑥+3𝑦 2 𝑥2 𝑥 𝑦 1
= 𝑒 [𝑥 𝑦 − − 𝑥𝑦 + + − ]
6 3 3 2 6

2𝑥+3𝑦
1 2 𝑥2 1 𝑥 𝑦 1
=𝑒 [ 𝑥 𝑦− − 𝑥𝑦 + + − ]
6 18 6 18 2 36
Non-Homogeneous linear partial differential equations with constant
coefficients
A linear partial differential with constant coefficients is known as non-
homogeneous l.p.d.e. with constant coefficients if the order of all the partial
derivatives involved in the equation are not all equal.
For example:
𝜕2 𝑧 𝜕𝑧
1) 2
+ +𝑧 =𝑥+𝑦
𝜕𝑥 𝜕𝑦
𝜕3 𝑧 𝜕2 𝑧 𝜕𝑧
2) + + = 𝑒 𝑥+𝑦
𝜕𝑥 3 𝜕𝑥𝜕𝑦 𝜕𝑦

Definition: A linear differential operator F (Dx, Dy) is known as (reducible), if it can

be written as the product of linear factors of the form aDx + bDy + c with a, b and c
as constants. F (Dx, Dy) is known as (irreducible), if it is not reducible.
For example, the operator 𝐷𝑥2 − 𝐷𝑦2 which can be written in the form (Dx - Dy) (Dx +
Dy) is reducible, whereas the operator 𝐷𝑥2 − 𝐷𝑦3 which cannot be decomposed into
linear factors is irreducible.
Note: A l.p.d.e. with constant coefficient F (Dx, Dy) z = f (x, y) is known as reducible,
if F (Dx, Dy) reducible, and is known as irreducible, if F (Dx, Dy) is irreducible.
OR
If in the equation
𝑓(𝐷𝑥 , 𝐷𝑦 )𝑧 = 𝐹(𝑥, 𝑦). . . . . . . . . (1)

The polynomial expression 𝑓(𝐷𝑥 , 𝐷𝑦 ) is not homogeneous, then equation (1) is a

non-homogeneous linear partial differential equation. As in the case of
homogeneous linear partial differential equations, its complete solution = C.F+P.I.
Note:
The methods to find P.I are the same as those for homogeneous linear equations.
Determination of Complementary function (C.F.) (the general solution) of a
reducible non-homogeneous linear partial differential equation with constant
coefficients
Let F (Dx, Dy) = (aDx + bDy + c)k, where a, b, c are constants and k is a natural number,
then the equation F (Dx, Dy) z = 0 will be (aDx + bDy + c)k z = 0 and the solution is
𝑐
𝑧 = 𝑒 −𝑎𝑥 𝜙(𝑎𝑦 − 𝑏𝑥) ; a ≠0, k = 1
Or
𝑐
𝑧 = 𝑒 −𝑏𝑦 𝜙(𝑎𝑦 − 𝑏𝑥) ; b ≠0, k = 1
For any k > 1, the solution is
𝑐
𝑧 = 𝑒 −𝑏𝑦 [𝜙1 (𝑎𝑦 − 𝑏𝑥) + 𝑥𝜙2 (𝑎𝑦 − 𝑏𝑥)+. . . +𝑥 𝑘−1 𝜙𝑘 (𝑎𝑦 − 𝑏𝑥)]; b ≠ 0
Or
𝑐
𝑧 = 𝑒 −𝑎𝑥 [𝜙1 (𝑎𝑦 − 𝑏𝑥) + 𝑥𝜙2 (𝑎𝑦 − 𝑏𝑥)+. . . +𝑥 𝑘−1 𝜙𝑘 (𝑎𝑦 − 𝑏𝑥)]; a ≠ 0
Where 𝜙1 ,…, 𝜙𝑛 are arbitrary functions.

Example:
Solve (Dx – 2 Dy +1)4 = 0
Solution: We have a = 1, b = -2, c = 1, k = 4
Then
1
𝑧 = 𝑒 −2𝑦 [𝜙1 (𝑦 + 2𝑥) + 𝑥𝜙2 (𝑦 + 2𝑥) + 𝑥 2 𝜙2 (𝑦 + 2𝑥) + 𝑥 2 𝜙3 (𝑦 + 2𝑥) +
𝑥 3 𝜙4 (𝑦 + 2𝑥)]
Where 𝜙1 ,…, 𝜙4 are arbitrary functions.
Example:
solve (𝟐𝐃𝐱
⏟ – 𝟑 𝐃𝐲 + 𝟏) (𝐃𝐱
⏟ + 𝟐 𝐃𝐲 − 𝟐) z = 0

linear linear
Solution: The given equation is reducible, then we have
a1 = 1, b1 = -3, c1 = 1, k1 = 1
1
− 𝑥
z1 = 𝑒 2 [𝜙1 (2𝑦 + 3𝑥)]
a2 = 1, b2 = 2, c2 = -2, k2 = 1
z2 = 𝑒 2𝑥 [𝜙2 (𝑦 − 2𝑥)]
The general solution is
1
−
𝑧 = 𝑧1 + 𝑧2 → 𝑧 = 𝑒 𝜙1 (2𝑦 + 3𝑥) + 𝑒 2𝑥 𝜙2 (𝑦 − 2𝑥)
2

Where 𝜙1 , 𝜙2 are two arbitrary functions.

Example:
Solve (𝑫𝟐𝒙 − 𝐃y) z = sin(x-2y)
Solution:
(1) The general solution z1 of (𝐷𝑥2 − Dy) z = 0 is
F (a, b) = a2 – b = 0 → 𝑎𝑖2 = bi
∞ 2
z1 = ∑ 𝐴𝑖 𝑒 𝑎𝑖𝑥+𝑎𝑖 𝑦
𝑖=1

(2) To find the P.I. of the given equation

1
P.I. = z = sin(𝑥 − 2𝑦)
𝐷𝑥2 −𝐷𝑦

a = 1, b = -2 → 𝐷𝑥2 = -a2 = -1
1
= sin(𝑥 − 2𝑦)
−1−𝐷𝑦
 1
Multiplying by
−1+𝐷𝑦

−1+𝐷𝑦
= sin(𝑥 − 2𝑦)
1−𝐷𝑦2

𝐷𝑦2 = -b2 = -4
−1+𝐷𝑦
= sin(𝑥 − 2𝑦)
1+4
1
= [−sin(𝑥 − 2𝑦) − 2cos(𝑥 − 2𝑦)]
5