Section 5.

5 Solving Nonhomogeneous Linear Differential Equations

In solving a linear non-homogeneous differential equation

a n y ( n )  a n 1 y ( n 1)    a 2 y ' ' a1 y ' a 0 y  f ( x)
or in operator notation,
Ly  f (x) ,
the right hand (forcing) function f(x) determines the method of solution. Fundamentally, the
general solution of this differential equation is
y  y p  yh
where y p is the particular solution to the original differential equation, that is,
Ly p  f (x)
and y h is the general solution to the homogeneous equation, meaning
Ly h  0 .
By the principle of superposition, we have
L ( y p  y h )  L( y p )  L ( y h )
 f ( x)  0
 f ( x)
It must be emphasized that we will always begin by finding the general solution of the
homogeneous case Ly = 0.

Differential Equations and their Operator Form

Differential Characteristic Linear Operator General Solution

Equation Eqn
r 2  3r  10  0 L  D 2  3D  10
y"3 y '10 y  0 y  c1e 5 x  c 2 e 2 x
(r  5)(r  2)  0  ( D  5)( D  2)
r 2  6r  9  0 L  D 2  6D  9
y"6 y '9 y  0 y  c1e 3 x  c 2 xe 3 x
(r  3)  0
2
 ( D  3) 2

y"16 y  0 r  16  0
2
L  D 2  16 y  c1 cos 4 x  c2 sin 4 x
y  c1 cos 3x  c 2 sin 3x 
y ( 4)  18 y"81 y  0 (r 2  9) 2  0 L  ( D 2  9) 2
x(c3 cos 3 x  c 4 sin 3 x)
r 4  r 3  12r 2  0
y ( 4)
y ( 3)
 12 y"  0 L  D 2 ( D  4)( D  3) y  c1  c 2 x  c3 e 4 x  c 4 e 3 x
r (r  4)(r  3)  0
2

The table of linear operators and solutions gives us a hint as to how to determine the
annihilator of a function. The functions that correspond to a factor of an operator are actually
annihilated by that operator factor.
How do we determine the annihilator?
Closely examine the following table of functions and their annihilators. Notice that the
annihilator of a linear combination of functions is the product of annihilators.

Table of Annihilators

f(x) Annihilator
a n x n  a n 1 x n 1    a1 x  a 0 D n 1
e kx Dk
x n e kx ( D  k ) n 1
cos bx or sin bx D2  b2
x n cos bx or x n sin bx ( D 2  b 2 ) n 1

( D  a) 2  b 2 
e ax cos bx or e ax sin bx
D 2  2aD  a 2  b 2

x e cos bx or x e sin bx
n ax n ax
( D  a)  b  
2 2 n 1

D  2aD  a  b 
2 2 2 n 1

3 x 2  2 x  5  7 e 2 x D 3 ( D  2)
3 x  5  7 xe 2 x  e 4 x D 2 ( D  2) 2 ( D  4)

The Annihilator Method

We can use the annihilator method if f and all of its derivatives are a finite set of linearly
independent functions. That is, f must be one of the following function types:
 Polynomial
 Sine or cosine
 Exponential (this includes hyperbolic sine and hyperbolic cosine)
 x n e kx , x n cos(kx) or x n sin(kx)
 A linear combination of the above.

If f(x) is of this form, we seek a differential annihilator of f, L f , so that L f ( f ) = 0. We

apply L f to both sides of the differential equation to obtain a new homogeneous equation
L f Ly  0 .
The Annihilator Method:
1. Write the differential equation in factored operator form.
2. Solve the homogeneous case Ly = 0.
3. Apply the annihilator of f(x) to both sides of the differential equation to obtain a
new homogeneous differential equation.
4. Identify the basic form of the solution to the new differential equation.
5. Determine the specific coefficients for the particular solution.
As a simple example, consider
3 y" y '2 y  2 x 2  6 x  4 .

Let us note that we expect the particular solution to be a quadratic polynomial.

Solution

We first rewrite the differential equation in operator form

(3D 2  D  2) y  2 x 2  6 x  4
and factor (if possible):
(3D  2)( D  1) y  2 x 2  6 x  4 .

We now identify the general solution to the homogeneous case (3D  2)( D  1) y  0 .
Since the characteristic equation is
(3r  2)(r  1)  0 ,
the roots are r = 1 and r   2 so the solution of the homogeneous equation is
3
2 x
y h  c1e 3
 c2 e x .

Now we identify the annihilator of the right side of the non-homogeneous equation:
D 3 ( x 2  5 x  7)  0

We apply the annihilator to both sides of the differential equation to obtain a new
homogeneous equation:
D 3 (3D  2)( D  1) y  D 3 (2 x 2  6 x  4)
giving
D 3 (3D  2)( D  1) y  0

The next step is critical because we must distinguish between the homogenous solution and
the particular solution to the original non-homogeneous case.
We know that the solution is (be careful of the subscripts)
y  yh  y p
 23 x
 c1e  c2e x  Ae0 x  Bxe0 x  Cx 2e0 x
2 x
 c1e 3  c2e x  
A 
 Cx
Bx 
2
  
yh yp

We must substitute y p  A  Bx  Cx 2 into the original differential equation to determine the

specific coefficients A, B, and C ( y 'p  B  2Cx and y "p  2C ):
3(2C )  ( B  2Cx)  2( A  Bx  Cx 2 )  2 x 2  6 x  4

Expanding and equating like terms gives

−2C = 2 (C = −1)

−2C −2B = 6 (B = −2)

6C − B − 2A = −4

giving A = 0, B = −2, and C = −1.

Therefore, the general solution to the original non-homogeneous equation is

2 x
y  c1e 3  c2e x  ( 2 x  x 2 ) (parentheses added for readability)
    
yh yp

Now consider
y"2 y '15 y  5 x  8e 3 x
Because the characteristic equation for the corresponding homogeneous equation is
r 2  2r  15  0 ,
we can write the differential equation in operator form as
( D 2  2 D  15) y  5 x  8e 3 x
which factors as
( D  5)( D  3) y  5 x  8e 3 x .

The characteristic roots are r = 5 and r = −3 of the homogeneous equation

( D  5)( D  3) y  0
correspond to the general homogeneous solution
y c  c1e 5 x  c 2 e 3 x .
The annihilator of the right-hand side 5 x  8e 3 x is D 2 ( D  3) because applying this operator
yields
D 2 ( D  3)(5 x  8e 3 x )  D 2 (5  15 x  24e 3 x  24e 3 x )
 D 2 (5  15 x)
00
0
Therefore, we apply D ( D  3) to both sides of the original differential equation to obtain
2

D 2 ( D  3)( D  5)( D  3) y  D 2 ( D  3)(5 x  8e 3 x )

0
We now solve the homogeneous equation
D 2 ( D  3)( D  5)( D  3) y  0 .

Again, we must be careful to distinguish between the factors that correspond to the particular
solution and the factors that correspond to the homogeneous solution.

We know that the solution is (be careful of the subscripts)

y  yh  y p
 c1e 5 x  c 2 e 3 x  Ae 0 x  Bxe 0 x  Ce 3 x
 c1e 5 x  c 2 e 3 x  
A 
Bx
Ce

3x
  
yh yp

We must substitute y p  A  Bx  Ce 3 x into the original differential equation to determine the

specific coefficients A, B, and C. (It is worth noting that Ce 3 x will only correspond to the
exponential term on the right side since it cannot contribute to the elimination of the other
terms. Thus, we have
9Ce 3 x  2( B  3Ce 3 x )  15( A  Bx  Ce 3 x )  5 x  8e 3 x

Expanding and equating like terms yields

(9C  6C  15C )e 3 x  15Bx  (2 B  15 A)  5 x  8e 3 x
which results in the equations
 15 B  5
 2 B  15 A  0
 12C  8
giving
2 1 2
A , B , C .
45 3 3
The general solution to the non-homogeneous equation is

2 1 2
y  c1e 5 x  c 2 e 3 x   x  e3x
   45 3
yh 3
yp
Special Case: When solutions to the homogeneous case
overlap with the particular solution
Let’s modify the previous example a little to consider the case when the solutions to the
homogeneous case overlap with the particular solution. Consider
y"2 y '15 y  5 x  8e 3 x .
Because the characteristic equation for the corresponding homogeneous equation is
r 2  2r  15  0 ,
we can write the differential equation in operator form as
( D 2  2 D  15) y  5 x  8e 3 x
which factors as
( D  5)( D  3) y  5 x  8e 3 x .

The characteristic roots r = 5 and r = −3 of the homogeneous equation

( D  5)( D  3) y  0
correspond to the general homogeneous solution
y c  c1e 5 x  c 2 e 3 x .

Again, the annihilator of the right-hand side 5 x  8e 3 x is D 2 ( D  3) .

We apply D 2 ( D  3) to both sides of the original differential equation to obtain

D 2 ( D  3)( D  5)( D  3) y  0
or combining repeated factors,
D 2 ( D  5)( D  3) 2 y  0 .

Note that the particular solution y p corresponds to the repeated factor D + 3 (since
e 3 x appears in the homogeneous solution) and the factor D2:
y p  Ax  B  Cxe 3 x .

We must now substitute y p and its derivatives into the original non-homogeneous differential
equation to determine the coefficients A,B, and C. We have
y p  Ax  B  Cxe 3 x
y ' p  A  3Cxe 3 x  Ce 3 x
y ' p  9Cxe 3 x  6Ce 3 x
so substitution yields
9Cxe 3 x  6Ce 3 x  2( A  3Cxe 3 x  Ce 3 x )  15( Ax  B  Cxe 3 x )  5 x  8e 3 x .

Combining like terms gives

 15 Ax  (2 A  15B)  (9C  6C  15C ) xe 3 x  (6C  2C )e 3 x  5 x  8e 3 x .
Equating the coefficients of like terms on each side gives
 15 A  5
2 A  15 B  0
8C  8
so that A = -1/3, B = -2/45, and C = 1. Thus, the general solution to the differential equation
is
1 2
y ( x)   x   xe 3 x  c1e 5 x  c 2 e 3 x .
3 45