Unit 16

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

1

UNIT 16 NON-HOM GENEOUS LINEAR


PARTIAL D FFERENTIAL
EQUATION WITH CONSTANT
COEFFICIE TS
Structure
16.1 Introduction
Objectives
16.2 Particular Integral
16,3 Analogues of Euler's Equations
16.4 Summary
16.5 Solutions/Answers

16.1 INTRODUCTION I

i
In Unit 15, we obtained the solutions of homog neous PDEs with constant coefficients. In
particular in Sub-sec. 15.2.2, we stated that for on-homogeneous partial differential
equation the complete integral consists of two rts, namely, complementary function
(which is the solution of the corresponding ho geneous equation) and particular integral.
In this unit, we shall develop methods of findin particular integrals for non- homogeneous
equations, that may be reducible or irreducible.

Objectives
After studying this unit, you should be able to
I
obtain particular integrals for reducible no PDEs;
obtain particular integrals for irreducible PDEs for cases when right
hand side is exponential function, sine or in x and y, and a
product of exponential function with
solve PDEs that are analogues of

Let us consider the following linear PDE of ord

We assume that Eqn. (I) is reducible, and then ow that a particular integral to it can be
found by treating successively m Lagrange equ Keeping in view that Eqn. (1) is
reducible, it can be expressed in the form

If we put (a2D+b2Dr+c2). . . . (amD+bmD'+cm) = u,


t . . . .(3)

t4
then the problem of solving Eqn.(2) is reduced t solving the Lagrange equation
@

(alD+bld'+cI) u = V(x,y) a . . . .(4)


The auxiliary equations corresponding to Eqn. ( ) are

The first two ratios in Eqn. (5) yields


I
where a, is an arbitrary constant. The first and the third ratios give Non-homogeneous Linear Partial
Differential Equations with
Constant Coefficients

Eqn. (6) is a first order linear ODE and its solution can be obtained by
Block 1. In the solution obtained if we replace a, by bl x-sly, we get
solution to Eqn. (4). We insert this expression of u in the linear PDE (
(m-1). The process may then be repeated for the new equation of the
continuing this way, we obtain the particular integral of Eqn. (1) in m
We shall assume that a complementary function of Eqn. (1) is known
in Sec.15.3 of Unit 15, and thus, a particular integral is to be found.
each Lagrange equation of the form (4), we may omit arbitrary func
obtained by the above method, general solution of Eqn.(l) can be w
combining the complementary function and the particular integral.
Let us illustrate this by the following example.

Example 1: Solve the equation

Solution : Eqn. (7) is a reducible equation because we may write it as


(D-2D') (D-D'- l )z = ( 2 + 4 ~ ) e - ~

One form of complementary function is


C.F. = f, (y+2x) + e-Y f2 (y+x),

where fl and f2 are arbitrary functions.

Note that the C.F. given by Eqn. (9) is one form of the C.F. You co
down the C.F. in this case as
C.F. = fl(y+2x) + e-xf2 (y+x) (Ref. Theorem 4, Unit 15)
To find a particular integral, let
(D-D'-1) z=u

(D-2D')u = (2+4x) e-Y

The subsidiary equations corresponding to Eqn. (1 1) are

dx-s!Y-
- - du
1 -2 (2+4~)e-~

1st and 2nd ratios yield,

and using this relation in 2nd and 3rd ratios, we get

du = (-l-2~)e-~dy

= (y-al-1) e-y dy

u = (al-y)e-Y + ar,? = + ar,?


Thus, the general integral of Lagrange Eqn. (1 1) can be expressed as
u = 2x e-)' + t$, (y+2x),
Second and Higher Order Partial where Q1 is an arbitrary function of y+2x.
Differential Equations
Since we know that an arbitrary function of y+2x, appears in the complementary function of
the given equation, we need not include it in solution (12). Consequently, we next solve the
equation

(D-D'-l)z = 2xe-r . . . . (13)


The subsidiary equations for Eqn. (13) are

&=*= dz
1 -1 z+2xe-Y

From 1st and 2nd ratios, we get

x + Y =pi.
pl being a constant.

From 2nd and 3rd ratios of Eqn. (14), we get

dz
-+z=-2xe-y
dY
= -2(p,-y) e-Y

This is linear ODE with eYas integrating factor. Its sclution is


zeY = -2ply + y2 + G2
= -y2-2xy + p2
Again, we may omit the arbitrary function, P2 = Q2 (y+x).

Consequently, a particular integral of Eqn. (8) can be expressed in the form

P.I. = (-y 2-2xy) e-Y = -(y2+2xy)e-Y

Thus, the complete integral of the given equation may be expressed as


z = fl(y+2x) + e-y f2(y+x)- (y2+2xy)e-Y

Remark : Variations in the process, such as change of the order of the factors in the
operator, may sometimes lead to different particular integrals. But these differences can be
very well taken care of by making suitable changes in the arbitrary functions which occ-ur in
the complementary function. For instance, in Example 1, if you change the order of factors
in Eqn. (8) then you will find that x2e-Yis also a particular integral. Now let us see how we
can change the solution obtained in Example 1 to a form which gives x2e-Yas a P.I.
We have

z = fl(y+2x) + e-Y f2(y+x)- (y2+2xy)e-Y

= fl(y+2x) + e-Y [ f2(y+x)- (y2+2xy)]


= fl(y+2x) + e-Y [ {f2(y+x)- ( y + ~ )+~x2
\ 1
= f l (y+2x) + e-Yg (y+x) + x2 e-Y

where g.(x+y) is a new arbitrary function equivalent to (f2(y+x)-(y+x)2 1.

This also provides an example for Theorem 3 of Unit 15 and reasserts the claim made there.

We can also use this method to solve an equation with variable coefficients if by a suitable
transformation of the variables, the equation can be reduced to an equation with constant
coefficients. For instance, consider the following PDE

( X ~ D ~ - ~ X ~ D D ' + ~ ~ ~=Dx3y4
'~+~~D')Z

Let u = Inx and v = Iny

Under this transformation of the independent variables, Em.( 15) reduces to


(D,(DI-I)-4DID1' + 4D1' (Dl'-1) + 6D1' ) z = e'U+4V Non-homogeneous Linear Partial
DitYerential Equations with
(Dl -2D1') (Dl-2DI'-I) z = e3U+4V . . . .(If) Constant Cwfficienb

a a
where D - -- and D ' - -
-au I -av
Now Eqn.(l6) has constant coefficients and can be treated by the method used in Example 1.

You may now try the following exercises.

E 1) Find the general integral of


z = eX+Y
(D~+DD'-~D'~)
E 2) Solve the equation
(D2+3DD'+2D") z = x+y

E 3) Solve the equation

( x ~ D ~ - z~ =~X D
Y ~ ~ )

It is not easy to find the particular integral of Eqn. (1) in the case when it i ; an irreducible
one. This is due to the fact that its reduction to the Lagrange equations carnot be effected.
- - a

Consequently, here we shall confine ourselves to some simple forms of thb function V(x,y)

F(D,D') z = V(x,y)
for which the equation can be solved easily.
As you go ahead, you will realise that the forms of V(x,y) which we shall consider here (viz.

I'
exponential, polynomial, sinusoidal etc.) are similar to the forms of non- rnogeneous terms
we considered while discussing the method of undetermined coefficients or solving ODES
(ref. Unit 6, Block 2). The reason for considering these particular forms

I same as discussed there.

We now discuss the methods applicable for the various forms of the
..
one. These methods are aoolicable both for reducible and
Exponential Function : Let us consider Eqn. ( l ) , when V(x,y) is an
of the form
V(x,y) = exp (ax+by)

where a,b are constants.

By differentiation, we know that


~r eax+by = eax+by

us eax+by -
- b s eax+by
D~DJS eax+by = ,r bs eax+by

:. F(D,D') eax+by= arsDr DjSeax+by

f
Operating upon both the sides of Eqn. (17), by the inverse operator F( ,D,) we get
9

1
I e"+b~ = F(a,b) -------
o n n'i
eax+b~
Second and Higher Order Partial If F(a.b) # 0, then dividirig both sides of the above equation by F(a,b), we get
Differential Equations

Rule corresponding to Eqn. (18) gives the P.I. in case V(x,y) is an exponential function. We
can also obtain this rule directly by using Theorem 8 of Unit 15.
Remember that Eqn. (18) gives the P.1. only when F(a,b) # 0. In case F(a,b) turns out to be
zero, it is often possible in such cases to make use of counterpart of the shift theorem for
PDE, i.e. Theorem 9 of Unit 15 to obtain the P.I. 1
If we have to solve i

then Eqn. (19) yields,


F(D,Df)z= F(D,Df) w eax+by= c eax+by

+ c eax+by= eax+byF(D+a,Dr+b)w(Using Theorem 9 of the Unit 15)


I
If a P.I. of this equation can be obtained then the required P.I. cah be written down using the t
relation z = w eax+by. i

I
We take up a few examples to illustrate the above theory.

Example 2 : Find a particular integral of the equation

(2D2-D!) z = 10e~-~Y

1
Solution : P.1. = ------ I O ~ ~ - ~ Y
~D~-D'

= 10 eXJY (Using rule (1 8))


2( 1124-3)

- 2ex-3~
Let us consider another example.

Example 3 : Solve (D2-Df2-3~+3~')z= ex-2~

Solution : The given equation is equivalent to

(D-D') (D+Dr-3) z = e x - 2 ~

:. C.F. = $,(y+x) + e3X42(y-~)


1 ,x-2y
P.I. =
(D-D') (D+D'-3)
1
.: f= (y+x) + e3X@2(y-x) --
12
eX-S
Non-homogemour Linar hrihl
Mfterentlnl E q u a t l o ~with
amstant Codklenb
We now take up an example in which F(a,b) corresponding to F(D,D') =
to be zero.
Example 4 : Find a particular integral of the equation
I
Solution :In this case F(D,DJ)= 2D2-D', a = 1, b = 2, c = 4 and F(a,b) = 0
I
Now, F(D+a, Dt+b) = 2 ( ~ + 1-) (Dt+2)
consideration reduces to
~
f
= 2D2+4D-D'. Thus Eqn. (20) in e case under

You can easily see that both w = x and w = - 4y will serve as a P.I.of this
Accordingly, the P.1.of Eq. (21) will be

You may now try the following exercise.


I
E 4) Solve the equations.
a) (D3-~DD'+D+1) z = e2x+3y
b) (D-D'-1) (D+D'-2) = e2X-y

We next take up the case when V(x,y) is a sine or a cosine function.


Sine or Cosine Function : You know that
exp [i (ax + py)] = cos (ax + py) + isin (ax+ by)
cos (ax + py) = Re exp [ i(ax + by)]
and sin (ax + py) = Im exp [ i(ax + py)]

That is, a sine or a cosine function can always be expressed as a combinacior of exponential
functions with imaginary exponents.
Thus, when V(x,y) is of the form
V(x,y) = cos (ax + fly) or sin (ax + By)
the methods discussed above can be used to obtain the P.I.of the given problem.
For the cosine function, the P.I.is
z = Re [exp [i(m + py)]/ F(ia, $)I, provided F(ia, ip) # 0.
For the sine function, the P.I. is
z = Im [exp Ii(ax + py)J/F(ia, $)I, provided F(ia, iJ3) # 0.

For instance, the P.I.of ( D ~ + ~ D D ' + D ' * - ~ D -=~sin(x+2y)


D~ can be obtaineci as
1 ei(x+2~)
P.I.= Irn
D~+~DD'+D'~-~D-~D'

-- I (6 cos (x+2y) - 9sin (x+2y)


- 117
Second and Higher Order Partial Alternatively, in case of non-homogeneous term being sine or cosine function it is straight
DiRerential Equations forward to use the method of undetermined coefficients. We set

z = Acos(ax + by) + Bsin (ax + by),


and substitute it in the given differential equation. Then equating the coefficients of
cos(ax + by) and sin (ax + by) on both the sides of given equation we determine A and B and
get the desired P.I. We shall illustrate the method with the help of an example. But before
that we would like to mention that the particular integral.in the case of sine or cosine
function can be obtained by still another method. You know, that

and DD' sin(ax+by) = -ab sin(ax+by)


Thus, in order to evaluate

replace D~ by (-a2), DD' by (-ab) and D ' ~by (-b2).

The similar arguments can be camed over for cosine function also.

We now take up a few examples to illustrate these methods.


Example 5 : Find a particular integral of the equation
(2D2-D') z = 8sin(x+2y)
Solution :Let z = Asin(x+2y),+ Bcos(x+2y)
where A and B are constants to be determined.

Substituting this value of z and its derivatives in the given equation, we get

-2Asin(x+2y) - 2Bcos(x+2y) - 2Acos(x+2y) + 2Bsin(x+2y) = 8sin(x+2y)


Equating coefficients of sin(x+2y) and cos(x+2y) on both sides, we get
-2A+2B = 8
*A=-2andB=2
-2B-2A = 0

:. P.I. = -2sin(x+2y) + 2cos(x+2y).


Let us consider another example.

Example 6 : Solve (D+D') (D+D'-2) z = sin(x+2y)


Solution : For the given PDE,

1
P.I. = sin(x+2y)
(D+D')(D+D'-2)
Non-homogeneous Linear Partial
Differential Equations with
Constant Coefficients

- -. 1 [ 6cos(x+2y) - 9sin(x+2y) ]
117

Hence, the general solution of the given equation is

z = Ol (y-x) + e2' Q2(y-x) + 1117 [ 6cos (x+2y) - 9sin(x+2y) ]


And now an exercise for you.

E 5 ) Solve the following partial differential equations.


a) (3DD1- 2 ~ ' ~ - D ' )z = sin(2x+3y)

b) (Dt2-DD'-2D)z = cos(3x+4y)

1!
We next take up the case when V(x,y) in Eqn. (1) is a polynomial in x an y.
Polynomial in x,y : Let us suppose that V(x,y) is a polynomial in x,y, o the form

where n is the degree of the polynomial. Now if the operator F(D,D') in


m, then we assume the P.I. to be a polynomial of degree (n+m), say z =
We then substitute the value of z and its derivatives in the given
like powers of x,y on both sides of the PDE to find the unknown

We now consider an example to illustrate the method.

Example 7 : Find the particular integral of the equation


I
Solution : Assume the polynomial
I

P
Substiiuting the values of z and its derivatives obtained from Eqn. (23) int Eqn. (22), we
obtain

4
Equating the coefficients of like powers of x and y on both sides of Eqn. ( 5 ) , we find

Eqns. (26) and (27) are two equations in three unknown constants.
Second and Higher Order Partial A simple choice for these constants satisfying both the equations is
Dimerentin1 Equations
= 0. a?, = I, a4," = 0

This leads to the following P.I.


',
z = x-y

Another choice of the constants, which satisfy Eqns. (26) and (27) is

This gives another P.I. as

1
An alternative way of obtaining the P.I. = V(x,y) when V(x,y) is a polynomial in x
F(D,D
and y is to expand -I - in an infinite series in ascending powers of D and Dr as
F(D,D1)
illustrated in the following examples.

Example 8 : Solve (DD'+D-D'-l )z = xy

Solution : The given equation can be re-written as

(D'+I) (D-l)z = xy

:. e: $(x) + ex ~ ( y )
C.F. = Y

I
P.I. =
DD' + D-D'- I

=.-[ 1-(D-D' + DD' ) I-' xy

= -[ I+D-D' + DD' + (D-D' + DD')' + . . . . I X Y


=-(xy+y-x+ I)

.'.z = e-Y$(x) + ex ~ ( y-)(xy+y-x+l )

Example 9 : Find the P.I. of (D'-a2D") z = x.


I
Solution : P.I. = X
~2 -a? ~ ' 2

You may note that P.I. in this case can also be obtained as follows
Nnn-homogeneous Linear Partial
Differential Equations with
Constant Coetlicients

I
Thus, P.I. obtained on expanding -in ascending powers of D is differ
F(D,D')
obtained on expanding --- in ascending power of D'. Since any P.I. is re
F(D,D')
obtain general solution, either of the above two methods can be used.

You may now try the following exercise.

E 6) Solve the differential equations


a) ( D ? + ~ D D ' + ~ Dz' ~=)x+y

b) (D2+2DD'+D'5 z = 12xy

c) (D2-6DD'+9Df2)z= 1 2x2 + 16xy

Many a time it may happen that in Eqn. (I), V(x,y) may be a product of
and another function of x and y (may be a polynomial in x and y or a
function). We shall therefore take up the method of finding P.I. for
functions.

Product functions : Let us suppose that V(x,y) in Eqn. ( 1 ) is a product of an exponential


function and a function of x and y, that is,

where a and b are constants.


' I
By Leibniz's theorem for successive differentiation, we have
I

+ ~ y = eax+by(D+b)Syf(x,y)
Similarly ~ ' " e ~ ~ yf(x,y)]
and DrDPS[eax+byyf(x,y)]= eaxfbY(D+a)' (D'+b)S yf(x,y)

= eax+by zxOC,
r s
(D+a)'(Dt+b)' yf(x,y) (using relation (30))

= eax+byF(D+a, D'+b) yf(x,y) . . . .(31)


Let F(D+a, D'+b) yf(x,y)= Q(x,y) . . . . (32)l

.. yf(x,y) = F(D+a,1D1+b)Q(x,Y) . .. . (33)


Substituting this value of yf(x,y) in Eqn. (31), we get
Second and Higher Order Partial 1
Differential 1;quatiuns Operating both sides of Eqn.(34) by the inverse operator F(D,D') ' we get

Thus, rule (35) gives the P.I. of Eqn. (1) in case V(x,y) is of the form (29).

Note that rule (35) is also useful while find& the

P. I. = ---- eax+by , and F(a,b) turns out to be zero. In such cases we regard @(x,y)= I in
F(D,D')
Eqn. (29), that is, we write
V(x,y)= e " x + b ~ = e " x + b ~ 1

Let us now consider the following examples.

Example 10 : Find the P.I. of


(D'+DD' +D+Dt+ I ) z = e-"(x2+y ')

1
Solution : P.I. = e-" (x2 + y2)
D' + DD'+D+Dt+ 1

- e-2x I (x2 + y2)


. (D-2)' + (D-2)D' + (D-2) + (D' + I)

+ higher order terns in D,D' ] ( x2 + y')

~ x a m p l e11 : Solve (D'-D" - 3D+3Dt)z = xy + eX+'y


Solution : The given equation can be expressed as

(D-D') (D+D'-3) = xy.+

C.F. = @,(y+x)+ e3' q2(x-y)


Non-homogeneous Linear Partial
Differential Equations with
Constant Coemients

Result given by Eqn. (35) could alternatively be achieved by assuming

z = exp(ax+by) zl (x,y) . . . . (36


Then

F(D,Df)z = exp (ax+by) $(x,y)

3 F(D,D') exp(ax+by) zl (x,y) = exp (ax+by) @(x.y)


3 exp(ax+by) F(D+a, D'+b) zl(x,y) = exp (ax+by) @(x,y)
(using Theorem 9 of Unit I!

3 F(D+a, D'+b) zl = $(x,y) . . . . . (3;


which gives the equation for zl.

P.I. of Eqn. (37) is

p.1. = 1
F(D+a,D'+b) $(x?Y)
1
and hence the required P.I. = exp(ax+by) - $ 6 , ~ (by
) using Eqn.
F(D+a,D'+b)

Let us now consider the following example.

Example 12 : Find a particular integral of the PDE


( D D ' ~+ D2 -2)z = 16 e2' cos3y

Solution : Assume the particular integral of the given PDE in the form
z = e2' w(x,y),

then
( D D ' ~+ D~-2) e2'w

= e2' [ (D+2) Dt2+ (D+212-2 ] w (using Theorem 9 of Unit 15)

= e 2 X [ ~ ~ ' 2 + 2 ~ ' 2 + ~ 2 +w2 ~ + 2 ]

Thus, the given equation is transformed to

( D D ' ~+ 2 ' +
~+ D2 ~ 2D + 2) w = 16 cos3y . . . . (3t
For finding the P.I. for Eqn. (38), let

w = Acos3y + Bsin3y.
where A and B are constants.

Substituting the value of w and its derivatives in Eqn. (38), we get,

2Acos3y + 2Bsin3y - 18Acos3y - 18Bsin3y = 16cos3y

Equating coefficients of cos3y and sin3y on both sides of the above equation

-16A = 16 and -16B = 0

*A=-l,B=O
Second and Higher Order Partial :. p.1. for reduced Eqn. (38) is
Differential Equations

Hence P.I. for the given equation is

And now some exercises for you.

E 7) Find the particular integral of

(D2-Df)z = V(x,y), when


a) V(x,y) = e3'+y

b) V(x.y) = e-'+Y (y+2x)

C) V(x,y) = e-3X+9~

E 8) Find the particular integral of


( D ~ - D ' ~z) = XY

E 9) Find a solution of

E 10) Solve the following PDEs.

a) (D~-DD' + D'- I ) z = cos(x+2y) + eY + xy + 1


b) ( 3 ~ ~ - 2 D ' ~ + D -zl )= 4 eX+Ycos(x+y)

C) (2D4+3D2 D' +Df2)z = 0

d) D(D-2D') (D+D') z = e X + 2(x2+4y2)


~

You may recall that while studying ODES with constant coefficients, we also studied Euler's
equations, which by means of certain transformations of the independent variables, could be
reduced to equations with constant coefficients. In PDEs also, there are equations with
variable coefficients which can be reduced to PDEs with constant coefficients and such
equations are, therefore, called analogues of Euler's equations. In the next section, we shall
first define such equations and then give transformations necessary to reduce them to PDEs
with constant coefficients.

16.3 ANALOGUES OF EULER'S EQUATIONS

Definition : A linear PDE with variable coefficients, characterised by the fact that the
I:
operator F(D,Df) consists of a sum of m s of the form

the a ' s being constant and which can be transformed to linear PDE with constant
coefficients is said to be an analogue of Euler's equation.

Examples of such equations are


(x2D' - y L ~ ' 2+ xD - yD') z = 0

(2x2D2-5xyDD' + 2y' D" +2xD+2yDr +4) z = 5xy2


( x 2 y ~ ' ~ ' - x y 2 ~ ~ ' ~ x 2 z~ =' 6x'/y
+ y ~ )

All equations falling under this category can be reduced to PDEs with constant cofrficients
1;.
by means of transformation of the independent variables given by Non-homogeneous Linear Partial
Differential Equations with
x = e U ,y = e V .. Constant Coefficients

Let us examine how the transformation defined by Eqn. (40) transforms


PDE with constant coefficients.
We have
I

and so on.
I
Hence,
I
Thus, each term of the form (39) becomes an expression involving power
constant coefficients only. In the case of non-homogeneous equations,
transformation is applied to the non-homogeneous term.
After the transformation of the independent variables
reduced to an equation with constant'coefficients,the
using methods given in Units 15 and 16. Once z has been found as a
can use the inverse transformation

u = Inx and v = Iny ....


to obtain a solution of the original equation in terms of x and y.

We illustrate this method by the following example.


I
Example 13 :Solve
D ~ + 2 y 2 ~ ' 2+ 2xD + 2yD')z = 5xy3
( ~ x -~5xyDD'
Solution : Using the transformation
x = e U a n dy = e V

the given differential equation can be expressed as


[ 2Du (Du-I) - 5 DUDv+ 2Dv (Dv-I) + 2Du + 2D, ] z = 5eue3v

3 (2Du2- 5DuDv + 2Dvz)z= 5e U + 3V

3 (2Du-Dv) (Du-2Dv)z = 5euf3v

C.F. = ( ~ v + u+) + 2 ( 2 ~ + ~ )

1 5 u+3v
P.I. =
(2Du -Dv) (Du-2Dv)
Second and Higher Order Partial Hence, z = $, (2u+v) + 4 2 ( ~ + 2 U+.)eu+3v
Dimerential Equations
Using inverse transformat~on

v = Iny and u = Inx.

we get

~ (xy2) I + $2 [1n(x2y) I + xy3


z = g , In

= yf,(xy2) + w2(x2y) + xy3.


And now an exercise for you.
- - -

E 11) Solve the following equations :

d) (6x2D2- xyDD - y 2 ~ ' 2+ 6xD - yD')z = 14xy cos (In(x/y)i

We now end this unit by giving a summary of it.

In this unit, we have covered the following.


1) Particular integral for reducible non-homogeneous linear PDE with constant coeffi-
cients can be obtained by repeated application of the Lagrange's method. The method
is repeated m times, if the equation considered is of order m.
2) For reducible and irreducible non-homogeneous equations, F(D,D')z = V(x,y), if
exp(ax + py
a ) V(x,y) = exp ( a x + py), then P.1. is
F(a.P)
*
), provided F(a, p ) O

b) V(x,y) = cos(ax + By). P.L is Re exp(i(ax+py)


F(ia. ip)
)]provided F(ia. ip) e 0

+JyH1
[ex-( i (ax
and if V(x,y) = sin(ax + by), P.1. is Im 4
F(ia.ip) 1.
provided F(ia, ip) # 0

Alternatively, assume z = Acos(ax + py) + Bsin(ax + py).

Substitute z and its derivative in the given DE and determine A and B by


comparing coefficients of cos(ax + py) and sin(ac + py) on both sides of the
given equation.

C) V(x.y) = Polynomial' in x and y, then assume


z= C a,, xrys,
r.s

substitute it in the given equation and by equating coefficients of like powers


of x and y on both sides, determine unknown constants a,, and obtain the
particular integral.

d) V(x.y) = exp ( a x + py). V I (x.y), then P.I. = e(ax+P~) I


F(D+a,D'+ p) v, (x.Y)
Alternatively, assume z = exp(ax + py). z; (x.y)

and given equation reduces to


For the cases (a) and (b) above if F(a, (3) = 0 and F(ia, ip) = 0, Non-homogeneous Linear Partial
Theorem 9 of Unit 15, P.I. can be obtained by any of the above Differential Equations with
Constant Coemcients
3) PDEs analogues of Euler's equations of ODE theory can be reduced
with constant coefficients by means of the transformation x = eUan

E 1) The given equation is


(D'+DD' - 2 ~ )z '= ex"
~

* (D+2D') (D-D')z = eX+Y

C.F. = 4(2x-y) +~(x+y)


To find P.I. let us put (D-D')z = u so that the given equation becom

(D+2DJ)u =

fi is a Lagrange's equation and its subsidiary equations are


I
The 1st and 2nd fractions, yield
2x-y = constant

The 3rd and 4th fractions, yield

=3 J ~ ( x + Yd(x+y)
) + constant
1
= - eX+Y+ constant
3

1
= - eX+Y+ function of (2x-y) (Using solution of 1st and 2nd
3

Since complementary function contains a function of (2x-y), we can o


function of (2x-y) in this P.I., and write the P.1. as

Then equation for z becomes the Lagrange's equation

1
I
(D--D') z = -
3

and 'its P.I. is

1
z = - x eX+Y
3

Hence, the general solution is


I
E 2) The given equation is
(D*+~DD'+ 2 ~ " )z = x+y

=, (D+D') (D+2D')z = x+ y
.: C.F. = @(x-y)+ ~(2x-y)
Second and Higher Order Partial 1
Differential Equations and P.1. = - (x+y)'
36

Hence the general solution of the given equation is

E 3) The given equation is


(x2D2- y 2 ~ ' 2 )=
z xy
Change the independent variables to u and v, where
u = Inx and v = Iny

Then given equation becomes

Now, C.F. = yf, (u+v)+ ~2 (u-V)


and P.I. = u eU+"

:. The general solution of the given equation is


z = yfI (u+v) + y2(u-v) + u eU+'

* z = $(xy) + yf(xly)+ (lnx) xy.

1
E 5) a) z = (x) + eS?$2 (2x+3y) + j cos (2x+3y)

1
b) z = $I (y) + e2XQ2 (x+y) + - [ cos(3x+4y) - 2sin(3x+4y)]
15

e3x+~
E 7) a) P.I. = --
8
b) The given equation is (D2-D')z = e-'+Y (y+2x)

We assume z = e-'+Y w(x,y), then

(D~-D') e-'+Y w (x,y)


= e-"+Y [ - (D' + 1) ] w (x,y)
-- e-x+Y [ D2-20-D'] w (x,y)
Then the given equation reduces to Non-homogeneous Linear Partial
DifTerential Equations with
Constant Coeflicients

Now assuming w = q, xrys and substituting in the above P ~ E .


r,s

we get

[ r (r-1) a,,, x'-~ y - 2r q,, x '-I yS- s q,, xryS-I 1 = y+2


r.s

Equating coefficients of like powers of x and y on both sides,

constants satisfying both the equations are

a],, =-2,a0,2=3/2,a2,, =a3,0=a2,0=0.


I
These are two equations'in five unknown constants. A simple ch ice of the

This gives the P.I. as

Hence the P.I. of the given equation is

C) The given equation is


I

Thus, we must treat e-3x+9Yas product of e-3x+9yand 1.


I
Second and Higher Order Partial
Differential Equations

Another way, we can find P.I. is as under :

1
P.I. = -
D2 xy2=-$[l +$lXy2
D'( 1 - 7)
D

I
E 8) P.I. = - x3y
6

E 9) The given equation is

a a
If D = - and D' = -, then the given equation can be written as (D~-D'~)z= xt
ax at

(D-D') (D+D')z = xt

C.F. = @(x+t)+ ~ ( x - t )

P.I. = --
6
x3t
:. Solution is z(x,t) = o(x+t) + ~(x-1)+ -
6

b) z = C ci eal +biy + -43 ex+ysin (x+y),


i=l
where 3aiz - 2b' + ai - 1 = 0

where ai, ci, di, and ei are arbitrary constants.

x4
C) z = o,(lnxy) + 49 (In -) + -3 sin (lny2)

d) z = o, ($11 +
[.
Y 4
[ ln (xy2)] - 1xy cos
[ :)
In-

You might also like