Annihilator Method

67
The Annihilator Method: An Alternative to Undetermined Coefficients

Introduction
In section 4.1 of our text, a method is presented for solving a differential equation of the
form
(1)
y' '+ py'+ qy = g (t ) .
where p and q are constants and g is some function of t. The method only works when g
is of a particular form, and by guessing a linear combination of such forms, it is possible
to determine the coefficients that make the proposed solution satisfy the equation. In
these notes I will present an alternative approach called the Annihilator Method.
Differential Operators. As a first step, we consider the differential operator D defined
by the equation Df(t) = f (t). With this notation, y = Dy and y = D2y. We can
combine these operations. For example, (D2 + 3D + 6) represents an operator that
transforms a function y to (D2 + 3D + 6)y = D2y + 3Dy + 6y = y + 3y + 6y. These
notational conventions allows us to rewrite equation (1)
( D 2 + pD + q ) y = g (t ) .
(2)
The expression in the parentheses is a polynomial in D, called the characteristic

polynomial of the differential equation. As we have already seen, the roots of the
characteristic polynomial give us solutions to the homogeneous equation. Lets look at
this in a new way, using operators.
Example: y' '3 y'4 y = 0 . In operator notation the equation is ( D 2 3D 4) y = 0. Now
factor the polynomial in D, to find ( D 4)( D + 1) y = 0. This shows that there are
eigenvalues of -1 and 4, so e-t and e4t are solutions to the differential equation. The new
way to look at this is to associate each solution with one factor of the polynomial. The
solution e-t goes with (D + 1) and (D + 1) e-t = 0. If we apply (D - 4) to both sides of the
equation we derive (D 4) (D + 1) e-t = (D 4)0 = 0. In this way we see that a solution
to the homogeneous equation for a single factor of the characteristic polynomial is also
a solution to the homogeneous equation for the full polynomial. Also, the factors of the
polynomial are interchangeable: (D 4) (D + 1) = (D + 1) (D 4), so we can repeat the
preceding argument with (D 4) and show that e4t is also a solution to the homogeneous
equation for the full polynomial.
With this new way of looking at things, we see that solutions to a homogeneous linear
differential equation can be found by factoring the characteristic polynomial in D and
finding homogeneous solutions for each factor. In general, the factor (D r) always
gives a solution ert because (D r) ert = 0. Here we adopt another new (and suggestive)
68
terminology. We say the operator (D r) annihilates ert because the operator

transforms ert to 0 (nothing).
Annihilating g(t). The annihilator method transforms an inhomogeneous equation to a
homogeneous equation by applying an operator on both sides. Here is an example:
Find the general solution of ( D 2 3D 4) y = 5e 2t . Notice that (D + 2) annihilates 5e-2t
because (D + 2) 5e-2t = 0. Applying (D + 2) to both sides of ( D 2 3D 4) y = 5e 2t gives
us ( D + 2)( D 2 3D 4) y = ( D + 2)5e 2t = 0 . Factoring the quadratic part we get
( D + 2)( D 4)( D + 1) y = 0. Now we have a homogeneous equation and know how to get
the general solution: y = c1e-2t + c2e4t + c3e-t .
Note that while every particular solution of the original equation must be of the form
just derived, the reverse is not true not everything of that form is a particular solution
of the original equation. For example, if we take c1 = c2 = 0 and c3 = 1, we get y = e-t,
and since that is a solution to the homogeneous equation ( D 2 3D 4) y = 0 , it cannot be
a solution to ( D 2 3D 4) y = 5e 2t . But we now know that if there is a particular
solution, it has to be expressible as y = c1e-2t + c2e4t + c3e-t with some specific choice of
the coefficients. Moreover, we might as well take c2 = c3 = 0 because c2e4t + c3e-t is the
general solution of the homogeneous equation, and so is annihilated by the differential
operator on the left side of the equation. That means it cannot contribute anything
toward our goal of producing 5e-2t on the right side. All that remains is c1. In this way,
we deduce that a particular solution MUST be of the form c1e-2t. We still have to find
c1, but we have determined that this is the only possibility, and have not had to guess.
Annihilators for various functions. What functions can be annihilated? Just those
that arise as solutions to homogeneous constant coefficient linear differential equations.
We have seen that functions of the form cert can be annihilated by D r. Working out
the details with complex eigenvalues we find that bet(cos t) and cet(sin t) are both
annihilated by (D2 2D + 2 + 2) = (D + i) (D i). We have also seen that
when r is a repeated eigenvalue, we get solutions of the form cert and ctert. Therefore,
ctert is annihilated by (D r)2. More generally, for any polynomial f (t) of degree m, the
function f (t)ert is annihilated by (D r)m+1 . This can be extended to the complex case,
or specialized to the case r = 0. On one hand, we find that the functions f (t)et(cos t)
and f (t)et(sin t) are annihilated by (D2 2D + 2 + 2)m+1 , and on the other hand
we find that f (t) itself is annihilated by D m+1 .
These results are summarized in the table below. For any multiple of a function in the
first column the appropriate annihilator is shown in the second column. You can also
use the annihilators for linear combinations. For example, to annihilate g(t) = 3t + 5e-2t,
you multiply the annihilators of the individual terms: D2 (D + 2). This works when the
individual annihilators have distinct eigenvalues. But suppose you want to annihilate
69
g(t) = te4t + 6t3e4t. For the first term the annihilator is (D 4)2 while for the second it is
(D 4)4. We dont have to multiply these, because one is a divisor of the other.
Instead we take the larger exponent, in this case 4. That is, since (D 4)4 annihilates
both te4t and 6t3e4t, it annihilates their sum.
Table of Annihilators
f(t)
Annihilator
a m t m + a m1t m1 + L + a1t + a0
D m+1
e rt
Dr
t m e rt
( D r ) m+1
cos t or sin t
D2 + 2
t m cos t or t m sin t
( D 2 + 2 ) m+1
et cos t or et sin t
( D ) 2 + 2 = D 2 2D + 2 + 2
t m et cos t or t m e t sin t
[( D )
+ 2
m +1
= D 2 2D + 2 + 2
m +1
Annihilator Method Algorithm

The table above gives you a complete catalog of functions that can be annihilated, and it
tells you what the annihilator is of each. How do we use that to find particular solutions
to a differential equation? Here are step by step instructions.
1. Express the differential equation in operator form h(D)y = g(t), where h(D) is a
polynomial.
2. Factor h(D) into linear factors for real eigenvalues and quadratic factors for
complex (nonreal) eigenvalues.
3. Form the general solution to the homogeneous equation h(D)y = 0.
4. Find an annihilator for g(t). Well call that k(D). It is a polynomial and will
already be in factored form.
5. Form the general solution to the homogeneous equation k(D)h(D)y = 0.
6. Remove from this second general solution the functions that already appeared in
the general solution of the original homogeneous equation.
7. Substitute the remaining general solution into the original equation h(D)y = g(t)
and solve for the unknown coefficients.
70
In the end, you still have to solve for the coefficients of the specified functions, just as
in the method of undetermined coefficients. In fact, the two methods work for exactly
the same types of equations, and use the same steps to find the coefficients. But the
annihilator method is better for two reasons. First, you know ahead of time exactly
what type of function g(t) must be for the method to succeed: any linear combination of
functions given in the first column of the table. Second, you dont have to guess the
correct form for the particular solution. That is generated automatically when you use
the annihilator method.
More Examples.
1. ( D 2 3D 4) y = 10e 4t . We factor the polynomial in D first, finding
( D 4)( D + 1) y = 10e 4t , and recognize that the general solution to the homogeneous
equation is y = c1e4t + c2e-t . Next, we recognize that the annihilator of
g(t) = 10e4t is (D 4). Applying this to both sides of the original differential
equation we get ( D 4) 2 ( D + 1) y = 0 . The general solution to this homogeneous
equation is y = c1e4t + c3 te4t + c2 e-t. We eliminate the terms we had before,
leaving y = c3 te4t . This is the form of the particular solution we need.
Substituting into the original differential equation, we have
(D2 3D 4) c3 te4t
c3((te4t) 3(te4t) 4(te4t))
c3((16te4t +8e4t) 3(4te4t +e4t) 4(te4t))
c3((16te4t 12te4t 4te4t) + (8e4t 3e4t))
c3(5e4t)
c3
=
=
=
=
=
=
10e4t
10e4t
10e4t
10e4t
10e4t
2
This gives us the particular solution ypart(t) = 2te4t, and the general solution to the
original equation is yhom(t) + ypart(t) = c1e4t + c2e-t + 2te4t.
2. ( D 2 3D 4) y = 3t 2 . We begin as before, by finding the general solution to the
homogeneous equation, y = c1e4t + c2e-t . Next, we recognize that the annihilator
of g(t) = 3t2 is D 3. Applying this to both sides of the original differential
equation we get D 3 ( D 4)( D + 1) y = 0 . The general solution to this homogeneous
equation is y = c1e4t + c2 e-t + c3+ c4t + c5t2. We again eliminate the terms in the
original homogeneous equation solution, leaving y = c3+ c4t + c5t2. This is the
form of the particular solution we need. Substituting into the original differential
equation, we have
71
(D2 3D 4)( c3+ c4t + c5t2)

( c3+ c4t + c5t2) 3( c3+ c4t + c5t2) 4( c3+ c4t + c5t2)
(2c5) 3(c4 +2 c5t) 4( c3+ c4t + c5t2)
( 2c5 3c4 4c3) + t(6c5 4c4 ) + t2( 4c5)
=
=
=
=
3t2
3t2
3t2
3t2
Equating coefficients of powers of t on each side gives us three equations:

2c5 3c4 4c3 = 0
6c5 4c4 = 0
4c5 = 3
and since this is a triangular linear system, it is not difficult to find the solution:
c5 = 3/4, c4 = 9/8, c3 = 39/32. This gives us the particular solution
ypart(t) = 39/32 + (9/8)t + (3/4)t2, and the general solution to the original
equation is yhom(t) + ypart(t) = c1e4t 39/32 + (9/8)t + (3/4)t2.
Exercises: For all the assigned problems in section 4.1, use the annihilator method to
find the form of the particular solution as a linear combination of specific functions.

Annihilator Method

Uploaded by

Copyright:

Available Formats

Annihilator Method

Uploaded by

Document Information

Original Description:

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

Annihilator Method

Uploaded by

Copyright:

Available Formats

67

The Annihilator Method: An Alternative to Undetermined Coefficients

The expression in the parentheses is a polynomial in D, called the characteristic

terminology. We say the operator (D r) annihilates ert because the operator

Annihilator Method Algorithm

(D2 3D 4)( c3+ c4t + c5t2)

Equating coefficients of powers of t on each side gives us three equations:

You might also like