Sunteți pe pagina 1din 18

Linear Algebra and Differential Equations

Lecture 18
Linear DEs with Constant Coefficients

July 16th, 2015

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
1 / 18

Constant Coefficient Linear Homogeneous DEs


A first order constant coefficient equation
x 0 = rx
has solution x = ce rt . This suggests that using this as an ansatz (fancy
word for educated guess in math) in constant coefficient DEs.
For example to solve the DE
x 00 + 4x 0 + 3 = 0
substituting the ansatz into the equation leads to
(r 2 + 4r + 3)e rt = 0

r = 1, 3

so the general solution is


x(t) = c1 e t + c2 e 3t

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
2 / 18

Distinct real roots


The general constant coefficient equation
x 00 + px 0 + qx = 0
Substituting x = e rt gives
r 2 + pr + q = 0
This is called the characteristic or auxiliary equation. This has two real
roots if p 2 4q > 0
r1,2 =


p
1
p p 2 4q
2

and then the solution is


x(t) = c1 e r1 t + c2 e r2 t

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
3 / 18

The two functions are linearly


rt
e1
e r2 t
W (t) =
r
t
r1 e 1 r2 e r2 t

independent since

p

= (r1 r2 )e (r1 +r2 )t = p 2 q e pt

which is non zero.


This generalizes to higher order equations.
distinct roots r1 , r2 , , rn then


e r1 t
e r2 t

r1 e r1 t
r2 e r2 t

W (t) =
..
..

.
.
(n1)
r1 t r (n1) e r2 t
r
e
1
2

If the auxiliary equation has n


e rn t
rn e rn t
..
.

..
.

(n1) rn t
e

rn

= (r1 r2 )(r1 r3 ) (rn1 rn )e (r1 +r2 ++rn )t


which is non zero, since the roots are distinct.

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
4 / 18

Complex roots
For second order equations, if the auxiliary equation
r 2 + pr + q = 0
satisfies p 2 4q < 0 then the roots are complex.
For example the DE
x 00 + 4x 0 + 13x = 0
has auxiliary equation
r 2 + 4r + 13 = 0
with roots
r1.2 = 2 3i
The general solution is
x(t) = c1 e (2+3i)t + c2 e (23i)t = e 2t (d1 cos(3t) + d2 sin(3t))
where the second expression is obtained using Eulers formula.

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
5 / 18

The solutions are linearly independent




e 2t cos(3t)
e 2t sin(3t)
W (t) = 2t
e (2 cos(3t) 3 sin(3t)) e 2t (2 sin(3t) + 3 cos(3t))

= 3e 4t 6= 0
For higher order equations if an equation has a complex root, its complex
conjugate must also be a root. As long as the real roots are distinct the
solutions will be linearly independent. The calculation of the Wronskian
follows just as in the distinct real roots case.
W (t) = (r1 r2 )(r1 r3 ) (rn1 rn )e (r1 +r2 ++rn )t 6= 0

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
6 / 18

Repeated roots
When we use the ansatz for the DE
x 00 + 4x 0 + 4x = 0
the auxiliary equation becomes
r 2 + 4r + 4 = 0

r = 2

and we dont have enough functions to span the null space which has to
be two dimensional. Now try a different ansatz: x = u(t)e 2t
x 0 = u 0 e 2t 2ue 2t

x 00 = u 00 e 2t 4u 0 e 2t + 4ue 2t

Substituting into the DE gives


u 00 = 0
Thus the general solution of the equation is
x(t) = (c1 + c2 t)e 2t

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
7 / 18

Multiple repeated roots


If the auxiliary equation has a repeated root of multiplicity m, r . This
corresponds to the DE

m
d
x(t) = 0
r
dt
Let x = u(t)e rt


Thus

d
r
dt

x(t) = u 0 e rt + rue rt rue rt = u 0 e rt

d
r
dt

m

x(t) = u (m) e rt = 0

with the solution of the original DE


x(t) = (c0 + c1 t + + cm1 t m1 )e rt

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
8 / 18

Cauchy-Euler equations
Cauchy-Euler equations have the form
t 2 x 00 + ptx 0 + q = 0
with p and q constants. In this case we use the ansatz : x = t r .
Substituting into the DE
t r (r (r 1) + pr + q) = 0
with roots



q
1
2
1 p + (p 1) 4q
r1,2 =
2
As long as these are distinct we have the general solution
x(t) = c1 t r1 + c2 t r2
If they arent distinct, reduction of order yields
x(t) = t r (c1 + c2 ln t)

Linear Algebra and Differential EquationsLecture 18 Linear DEs


Julywith
16th,
Constant
2015 Coefficients
9 / 18

An Example
Find the general solution to the DE
t 2 x 00 + 3tx 0 + 10x = 0
The auxiliary equation is
r 2 + 2r + 10 = 0
with roots
r1,2 = 1 3i
and general solution
x(t) = e t (c1 t 3it + c2 t 3it )
but
t 3i = e 3i ln t
Thus the solution can be written as
x(t) = e t (d1 cos(3 ln t) + d2 sin(3 ln t))

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
10 / 18

Nonhomogeneous equations
Since the constant coefficient equations are relatively simple,
nonhomogeneous equations can be solved as long as the inhomogeneity is
reasonably simple also.
For example to solve the IVP
x 00 + 4x 0 + 8x = 8 sin(4t)

x(0) = 1

x 0 (0) = 2

First we need to find the solution of the homogeneous equation. The


auxiliary equation is
r 2 + 4r + 8 = 0 r = 2 2i
and the homogeneous solution is
xh = e 2t (C1 cos 2t + C2 sin 2t)
Derivatives of the inhomogeneous term involve only multiples of sin(4t)
and cos(4t), so for a particular integral pick a linear combination of these
xp = A sin(4t) + B cos(4t)

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
11 / 18

Substituting into the DE and gathering like terms together


sin(4t)(8A 16B 8) + cos(4t)(8B + 16A) = 0
Since sin(4t) and cos(4t) are linearly independent, the terms in
parentheses have to be zero. Solving gives
A=

1
5

B=

2
5

The general solution is


2
1
x(t) = sin(4t) cos(4t) + e 2t (c1 cos(2t) + c2 sin(2t))
5
5
The initial conditions give
c1 =

6
5

c2 =

3
5

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
12 / 18

Another example
Find the general solution to
x 00 + 4x = 12 cos(2t)
To find the solution to the homogeneous equation, the auxiliary equation is
r2 + 4 = 0

r = 2i

So the solution to the homogeneous equation is


xh = C1 cos 2t + C2 sin 2t
The problem is that trying a solution like the previous problem wont work
as the trial solution since they are in the null space of the operator.
Instead we will try
xp = t(A cos 2t + B sin 2t)

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
13 / 18

xp0 = t(2A sin 2t + 2B cos 2t) + A cos 2t + B sin 2t


xp00 = t(4A cos 2t 4B sin 2t) + 2(2A sin 2t + 2B cos 2t)
then
xp00 + 4xp = 2(2A sin 2t + 2B cos 2t) = 12 cos 2t
Thus A = 0, B = 3 and the general solution is
x = xp + xh = 3t sin 2t + C1 cos 2t + C2 sin 2t

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
14 / 18

Annihilators
Define the operator
df
dt
then a constant coefficient linear homogeneous DE can be written as
Df =

p(D)x = 0
where p(D) is a polynomial in D. A polynomial in D is an annihilator of a
function f (t) if f (t) is in the null space of the operator.
Some examples
(D + r )e rt = 0
(D+r )n+1 t n e rt = 0

(D2 2aD + a2 + b 2 )e at (c cos bt + d sin bt) = 0


(D2 2aD+a2 +b 2 )n+1 t n e at (c cos bt+d sin bt) = 0

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
15 / 18

Using annihilators to solve inhomogeneous equations


Given an inhomogeneous equation
P(D)x = f (t)
then if Q(D) is an annihilator of f (t) then x(t) must satisfy the
homogeneous equation
Q(D)P(D)x(t) = 0
This means any particular solution of the first DE must be in the null
space of the higher order DE. Thus we can use members of the null space
of the homogeneous equation that are not in the null space of P(D) as
trial functions for the particular solution.

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
16 / 18

An example
Find the general solution to the DE
x 00 3x 0 + 4x = 50te t
The solution to the homogeneous equation is
r 2 3r 4 = 0

r = 4, 1

xh = C1 e 4t + C2 e t

An annihilator for te t is (D+ 1)2 thus the particula solution will be a


solution of the homogeneous equation
(D + 1)3 (D 4)x = 0
x = C1 e 4t + e t (C2 + C3 t + C4 t 2 )
so our trial solution for xp will be
xp = e t (C3 t + C4 t 2 )

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
17 / 18

xp0 = (C3 t + C4 t 2 )e t + (C3 + 2C4 t)e t


xp00 = (C3 t + C4 t 2 )e t 2(C3 + 2C4 t)e t + 2C4 e t ]
xp00 3xp0 4xp = 5(C3 + 2C4 t)e t + 2C4 e t
Thus
50t = 5(C3 + 2C4 t) + 2C4
5C3 + 2C4 = 0
C3 = 2

50 = 10C4
C4 = 5

and the general solution is


x = (2t 5t 2 )e t + C1 e 4t + C2 e t

Linear Algebra and Differential EquationsLecture 18 Linear DEs


July with
16th,Constant
2015
Coefficients
18 / 18

S-ar putea să vă placă și