WITH CONSTANT COEFFICIENTS
53
LINEAR D.E.
2.3 Inverse differential operator and the particular integral
d
We have the differential operator D =
dx
Letussuppose that D[G (x)] =
F(x)
Then F ( x ) dx = G(x) is an inverse operation.
Equivalently we writeF(x) =G(x)
Theoperatorthat
D
stands for the integral is called an inverse differential operator.
1 1
etc. stands for successive integration.
D2 D
3
Examples: = [?át - COs X
=
sin x
f e s r
Note: We omit the constant ofintegration in order to have a specific expression.
Next, if 'a' is a constant let us find F(x)
y then (D-a) y =F(x) dy ay =F (x) which is a
Suppose F (x) = or
dx
linear differential equation whose solution is given by
yea =
F (x)eadx or y =
F (x)e dx
D-F(x) ] F ( x ) " d x ..(1)
D-a
1
Also (x)
D.-bD- F(x) (2)
=
(D-b)(D-a)
This is nothing but the repeated application of the result (1).
1 A B
Note: Alternatively cAn be expressed in the form D -b D-a by
(D-b) (D-a)
partial fractions so that
1 )and (1) can beapplied
(D-b)(D-a)F(x) = A-EFO
independemtly.
54 DIFFERENTIAL EQUATIONS2
Suppose (D-b) (D-a)
1 F(x) =
y then (D-b) (D-a)y
=
F(x)
In general we can say that,
F(X)
fD)y =
F(x) y = fD)
F(x) then y y, (x) free from the
=
But we havealready said that if f(D)y =
arbitrary constants is called the particular solution
or particular integral (P.)
denoted by y, Thus we can say that
.
F(x)
P.I y» f(D)
in respect of the equation in the form f(D) y = F(x)
and b
Furtheritmayalso be noted thatif F(x) aFj (x) +bFz (x), where = a are
constants, then
Fx) 5*)F2(x)
P.I Yp +f(D)
=
f(D)"f(D) "fD)
We present a few illustrations of finding P.I by this basic method.
1
(i) Let us find D-1 COs
cos 2x
2x
1
Using F(x) =
Je F (x) dx we have
D-T COs 2x = " J* cos 2 x dx .. (1)
Using the standard integral,
cosbx dx =2+b2 a
J" cosbxdx = cos bx + b sin bx ) in the R.HS of (1) we get,
coscos2x = ( - c o s 2x +2 sin 2 x)
D-1
ie. D-TCOs 2x = ( 2sin 2x-cos 2x)
-1
1
(ii) Let us find e
D +3D +2
1 1
Here
D+3D+2 (D+1)1) (D+2) by by partial fractions.
LINEAR D.E. WITH CONSTANT COEFFICIENTS
55
1
- 1
D +3D +2 D+2
1
Thus
D+3D+2 6
This fundamental method becomes highly tedious when f(D) is in higher orders
and F(x) is in a complicated form as we haveto necessarily decompose f(D) into
theform (D-m,) (D-m,) (D-m,) forapplying theresult.Henceweproceed
to illustrate simple and shortcut methods for finding the P.I when F(x) in
f(D)y = F(x) belong tosome specific form.
We discuss five specific forms of F(x)
-- - - - -
2.4 Specific forms of the particular integral
2.41 Type-1: P.I
of the form f(D)
f(D)
We have to obtain the particular solution of the d.e f(D)y = e
Let fD) =D +a, D+a2
We have D) =
ae, D() =X
.
fD) = (D+a, D + a,) = D()+a, D (M) +a,e*
a+a, a +a,e
=
(a +a, a +a, ) = f(a)
Thus f(D)A* =f(a) *
Operating with 1/f(D) on both sides we get,
=f(a) F(D) or
Sa)
Thus (Dbeing replaced by a) where f(a) *00
D)f(a)
Suppose f(a) = 0 then (D-a) is a factor of f(D)
56 DIFFERENTIAL EQUATIONS 2
e. f(D) = (D-a) ¢(D) (1)
1
f(D) (D-a) o(D) D-a ¢(a)
1
1
(a) (D-a (a)
1
o(ay1dr =1 2
f(D) (a)
D variable.
(D -a)o'(D) +¢(D) by treating
as a
From(1) f'(D) =
f' (a) =
0+0 (a) or f' (a) ¢ (a) =
Thus (2) becomes
where f(a) = 0 and f' (a) * 0
fD)*fta)
This result can be extended further also.
er
and so on.
Thatis, if f'(a) =0, f(D) *f" (a)
=
Note: The result also holds good for the following functions.
(ii) any constant k write k = k.e*
(i) e T as we can
(ii) a =
( Oga F = x where b =
log a
(io) sin h ax = 1/2 (e - e ) and cosh ax = 1/2 ( " + e " )
Illustrations
1. D+3 2+3 5
2. 6
D+3D+2 12+3.1 +2
[Refer the previous article, as we have found P.I by the basic method]
p3x +4 eSr+4
e t +4
3. 40
D+D+D+1 (3+ (3 ? +3+1
LINEAR D.E. WITH CONSTANT COEFFICIENTS
57
4 (eog4 og 4 x
D+ 5D+6 D+5D+6 D+5D +6
og 4 x
4
(log 4) + 5 (log 4) + 6 (log 4 ) +5 log 4 +6
e
(Dr. =
00)
D+7D+6 (-1)f+7(-1)+6 1-7+6
e
= X
2D+7 2(-1)+7 *5
e 2r -2x
6.
2x (Dr. = 0)
D+4D+ 4 (-2 )+4 (-2) +4 4-8 +4
e2
-2x
XD +4
XD-=X*
-4+4
A (Again Dr. = 0)
2x e 2r
= ie, 2
D+4D+4
cosh 2x 2 12x
1. 2 D-4 2p?-4 Ph +P2
D-4 D2-4
e 1
(Dr. = 00)
= 1e2* 2x
(Dr. = 0)
P2 D-42 (-24
-2x t e 2 r
-8
e t
x sinh 2x
P.I =
Pt+ P2 8 +8 4
58 DIFFERENTIAL EQUATIONS 2
sin ax COs ax
2.42 Type-2 P.i ofthe form
f(D) SD)
D ( sin ax) = a cosax D° ( sin ax) =-asin.ax
D' (sin ax) = - a' cos a x, D* (sin ax) = a" sin ax
We observe that (D)' sin axr = (-a)' sin ax,
(D sinax = (-a# sin a x etc (DÉY" sin ax = (-a#" sin ax
Hence, f( D) sinax =
f(-af) sinax
Operating with 1/f(DÉ) on both sides we have,
Sin ax sin ax
sinax f(-a2). =
sinax or
where f- a) #0.
fD) * f(D) f(-a)
(D being replaced by - a)
It can be easily seen that the working does hold good for cos ax also.
COS ax COS ax
fD) f-2)
where f(-#)+0
However if f(-a^ ) =0, as in the previous type we can show that
Sin ax
2, Sin ax cosax COS aX
Note:
D)f(-) f(D)f'(-?) etc.
(1) The results also hold good for the functions sin (ax +b) and cos (ax +b)
(2) Functions like sin axcos bx, sin ax sin bx, cos ax cos bx have to be converted into a
sum
by using basic trigonometric formulae.
(3) After replacing D by -a,
f(D) transforms into the form aDtß (a & B ae
constants) in which case we multiply both the numerator and
denominator
aD+ B for
proceeding further in obtaining the P.I. However if f( D) involres onlyby
even
powers of D then, the same transforms into a constant on
replacing Dby a -
Illustrations
1. COs 2x
1.
1D-1 We need to multiply and divideby (D+1)
P.I= (D+1)cos 2x D cos 2x
+cos 21 -2sin 2x + cos 2x
(D+1)D-1) D-1 D-1
Here a =
2 and we replace D° by --4
LINEAR D.E. WiTH CONSTANT COEFFICIENTS 59
P.I =
-2
sin2x + cos 2x
(2 sin 2 x -cos 2 x)
-4-1
IRefer (i) in the article 2.3 as we have found P.I by the basic method]
sin 3x
2 D+
4D +3
Here a= 3, replace D by -9
sin 3x Sin3x 4D+6.
P.I
-9+4D +3 4D -6
Multiply and divideby
(4D +6) sin 3x 4 (3cos 3x) +6 sin 3x
(4D +6) ( 4D-6) 16D-36
PI 6 (2 cos 3x +sin 3x)_6(2OS 3x + Sin 3x) =2(2 cos 3x + sin 3x )
16 (-9)-36 180 B0
2.43 Type-3: P.I ofthe form 9 ) where o (x) is a polynomial in x
f(D)
We are seeking the particularsolution of f(D)y = o (*) where
(x) =a + a x t . t a , -1 Xt4
It is evident that y = y, will also be a polynomialin x. Hence P.I is found by division.
P
By writing o (x) in descending powers of x and f(D) in ascending powers of D,
the division geis completed without any remainder. The quotient so obtained in the
process of division will be the particular integral.
Illustrations
1. 2x+4x+11
D+3D +2
2 + 4 x+1 and we shall divide.
P.I =
2+ 3D + D
2+ 3D +D 2x+ 4x+
2x+6x +2 Note:3D (a) =6r ; D (1)=2
-2x-1
- 2x 3
Note:3D(-x) =-3, D*(-x) =0
(4)(+)
+2
2
DIFFERENTIALL EQUATIONS
60
P-l x-x+ 1
Thus
Thus =
=
2x + 4x +t1
2)(x-X+1)
easily seen that, (LD + 3D +
Remark : Itcan be
-3x +1 3x +for division
2 P - D+D
D-D
First we need
-/12 + / 6
-D+D 3x+1
D
-dx dx= 12
2x
( (+) Also D° (-x/12 ) =-2x
Now we need
+1
r+1
(+) ( - D
xdx dx =
0 Also D' (x'/6) = 1
Thus P- 12(2-x)
e V function of x
the form where Vis a
2.44 Type-4: P.I of f(D)
Let W be any function of x.
W
D( W) =eDW +a e W ="(D +a)
Operating with D again we have,
W
D(W) = D(D +a) W+D ( " ) (D+a)
= ( D + aD ) W +a e (DW+ aW)
= ( D W+ 24 DW+a W)
D (W) = ( D +a) W and so on
W .(1
f(D)( W) "f(D +a) =
V. Hence W V
Let f(D+a) W =
f(D+a)
Thus (1) becomes
V =V
f(D)f(D+a)
S(D)*
Now operating with 1/f( D) on both sides we have,
INEAR D.E. WITH CONSTANT COEFFICIENTS
61
fD+a) fD)
Thus f(D)
V
fD+a)
Remark: 1f " is multiplied with somne function of x, need to
Then the coniputation of P.I will follow the
we
first shift D to D +a.
type 2 or 3.
Illustrations
sin 3Bx,
1.
D-1
Shifting D to (D +2) we have,
sin 3x sin 3x
e
(D+2-1 D+4D+3
sin 3x
We have to find which is same as illustration-2 in Type-2
D+ 4D+3
Thus esin 3x 2x (2 cos 3x +sin 3x ) =F
(2 cos 3x + sin 3x)
D-1 30
2 (2x +4x +1)
D+5D+6
Shifting D to (D-1) we have
e
2x+4x+1 -x 2x+4x +1
(D -1 +5 (D-1)+6 D +3D +2
P.Iof 2+4r+1 is same as illustration (1) of Type-3.
D +3D +2
Thus (2x +4x +1) * (-x+ 1)
(D+5D+6)
V
2.45 Type-5: P.I ofthe form D where Vis a function of x
S(D) 'f(D)
Let W be any function of x.
D" (x W) = xD" W+n 1 D"- W, by Leibnitz theorem.
d D") by
Weshallwrite n D"- present in the R.H.S equal to regarding Das a
variable.
62 DIFFERENTIAL EQUATIONS
a
D(W) x (D" W) + (D") W
lence f(D) tW =
rf(D) W+f' ( D) W
V and hence W= V
Now, let f(D) W =
f(D)
(1) becomes
1 V 1
=XV+f' (D)fD)
fDF(D)
Operating with 1/f(D) on both sides we have
1
V
f(D) f(D)""(D) f(D)
7f(D) v f(D)
AFCD)(D)
f'D) xV
fD) f(D) f(D)
Thus f(D) f'(D)_V
f(D) fD)
Note
1. By repeated application of this result we can find the particular integral of the furnrtios
V, V etc.
2. In the cases of " cos ax and x" sin ax specially when n> 1 we prefer thefollowr
alternative method
We know that e"= cOSax +i sinax
COS X= Real part (R.P) of ex
Sin ax = Imaginary part (I.P) of ex
Hence x" cosax = R.P ( e"""x") and "sina= I.P (e )
is done by first shifting D to D+ ia and then have to divide
Computation of P.I ne
b y the transformed form of f( D). By separating this into real and imaginary pars
we obtain the P.I of x" cos axf( D) and " sin avf ( D) respectively.
LINEAR D.E. WITH CONSTANT COEFFICIENTS
63
Illustrations
X COS X
1.
D+9
Here V= cos x and f(D) D+9
=
'. f' (D) =
2D
cOs x 2D cosx
Now
D49 D +9 D+9
2D cos X
D by
D+9 8 byreplacing -1
sin x x COS+ 1+9)
8
. Sin X
x COS X sin x
4 (D+9) 8 8 32
Thus COs X
D+9 32*cOs X+sinx)
Aliter: X COSX
R.p X R.p
D+9 D+9 ?+9
D+9
X
12 = R.P e"
(D+i)+9 R.Pe D+2iD +8
= R.Pe*.
= -1)
We divide x by 8+2iD+ D
/8-i/32
8+2iD+D
x+i/4
Quotient = A)
(-)
- i/4
-i/4
(+)
P.I =
R.P*(4x -i)
R P{ ( cos x + i sin x ) ( 4r- i)
1
75R P( 4x cos x + sin x) + i ( 4xsinx - cos x )
32
P.I = ( 4x cos x + sin x )
64 DIFFERENTIAL EQUATIONS
2.5 Summary of all the results
Solution of f(D) y=¢(x)
Stage-1 Stage-2
Finding CF (y,) by Finding P.I ( y,) where
solving f(D)y 0 y 9 (x) /f(D)
Complete solution: y=C.F+ P.I or
y=y,+ y,
Stage-1: Solution of f(D) y = 0
Form the A.E f(m ) = 0 and solve.
C.F is based on the nature of the roots of the A.E as summarized in the following
table.
Nature of the roots of the A.E Complimentary Function (C.F)
me
1. 11,l, .
"l c, e"1* + C% e"2* + .+ C.
(real and distinct)
2. 1, m, =* =m, m (C +C2 X+ Cg++ C, x"-) x
*
(1 Coincident real roots)
3. (a) ptiq P" (C, cos
eP
qx+ C, sin qx)
(a pair of complex roots) C, cosqx + C, singx
(b) tiq
(a pair of imaginary roots)
times
C+C2 X+ +C ) cos qx
(C) ptiq repeated n
+(C +x+. +61 x-) sin qx
LINEAR D.E. WITH CONSTANT COEFFICIENTS 65
Stage-2: Particular solution of f(D) y = o(x) or P.I = y, = ¢ (x Vf(D)
Type of P.I Method of getting the P.I
1. e
e
a)z0;
f(D) f(a)
(Also applicable for (D is replaced by a
a , a = e l o ga
1x x
e elt
e
sinhax coshax by using their etc.
definitions) f'(a) f"(a)
[iff(a) =
01, [if f' (a) =
0
Sinax COS axX Sinax COSax
where f(-a) # 0
f(D) f(D) f(-) f-a)
(Also applicable for (D is replaced by - )
sin (ax +b), cos ( ax +b)& Sin ax COS axK
sinax cos bx, sin ax sin bx X
cos ax cos bx by changing into f'-#) ff(-)
if f(-a) = 0 etc.
sum)
Divide (x) byf( D) by writing ¢ (x) in
|3. ( where o (x) is a
fCD) descending powers of x and f( D) in
polynomial in x. ascending powers of D. Quotient =P.I
as remainder will be zero.
V
4 xVV e
f(D)
where V = V(x)
f(D+a)
Dis replaced by D+a first & then find
V/f(D +a) premultiplied by e
V
5. xV T-D)
;V = V(x) fD)Jf(D)
f(D)
V xV (Repeated application),
f(D) fD) Lp e "
x cosax sin a* R.P e I.P.
f(D)' f(D)
f(D) f(D)
