Tuan 11.12 InhomogeneouslinearODE

Chapter 5: Second order linear ODEs (continued)
Nguyen Thu Huong
School of Applied Mathematics and Informatics

Hanoi University of Science and Technology
January 9, 2023
Inhomogeneous second order linear equations
Content
2. Inhomogeneous second order linear equations

2.1 Variation of parameters
2.2 Inhomogeneous equations with constant coefficients
2.3 Euler equations
2.4 Mass – spring system: forced oscillation
2nd order linear DEs Nguyen Thu Huong

Recall
Theorem (Superposition of solutions)
Assume that
y1 is a solution of the equation y 00 + p(x)y 0 + q(x)y = f1 (x),
y2 is a solution of the equation y 00 + p(x)y 0 + q(x)y = f2 (x).
Then y1 + y2 solves the equation
y 00 + p(x)y 0 + q(x)y = f1 (x) + f2 (x).

Recall
Theorem
The general solution to the inhomogeneous equation
y 00 + p(x)y 0 + q(x)y = f (x) (1)
has the form y = ȳ + y ∗ , where

ȳ is the general solution to the corresponding homogeneous
equation y 00 + p(x)y 0 + q(x)y = 0,
y ∗ is a particular solution to (1).
Last week: Solve for ȳ = C1 y1 (x) + C2 y2 (x). In the special case of

an ODE with constant coefficients: characteristic equation.
Today: We aim at y ∗ .

Content

2.3 Euler equations
Variation of parameters
The general solution of the corresponding homogeneous equation
y 00 + p(x)y 0 + q(x)y = 0 is:
ȳ (x) = C1 y1 (x) + C2 y2 (x).
We look for a particular solution of the inhomogeneous equation in

the form:
y ∗ (x) = C1 (x)y1 (x) + C2 (x)y2 (x)
Substituting in the equation
y ∗ = C1 y1 + C2 y2
(y ∗ )0 = C1 y10 + C2 y20 + C10 y1 + C20 y2
| {z }
0
∗ 00
(y ) = C1 y100 + C2 y200 + C10 y10 + C20 y20 .
Hence Ly ∗ = C1 Ly1 + C2 Ly2 + C10 y10 + C20 y20 ⇒ C10 y10 + C20 y20 = f (x).
A particular solution of the inhomogeneous equation is found in

the form:
y ∗ (x) = C1 (x)y1 (x) + C2 (x)y2 (x),
where
( C10 (x), C20 (x) satisfy the system
C1 (x)y1 (x) + C20 (x)y2 (x) = 0,
0
C10 (x)y10 (x) + C20 (x)y20 (x) = f (x).

(
C1 (x) = ϕ1 (x) + K1 ,
We obtain
C2 (x) = ϕ2 (x) + K2 , K1 , K2 ∈ R,
The general solution is given by
y = (ϕ1 (x) + K1 ) y1 (x) + (ϕ2 (x) + K2 ) y2 (x).

How to solve y 00 + p(x)y 0 + q(x)y = f (x)
1 Solve the homogeneous equation y 00 + p(x)y 0 + q(x)y = 0. If

a solution y1 is given, we find another one y2 by Liouville
formula. Then we obtain ȳ = C1 y1 + C2 y2 .
2 Find a particular solution of the inhomogeneous by the method
of variation of parameters: y ∗ = C1 (x)y1 (x) + C2 (x)y2 (x).

Example
Solve the following equations
x 1
a) y 00 + y0 − y = x − 1 that the corresponding
1−x 1−x
homogeneous equation has a solution y1 = e x .
e −x
b) y 00 + 2y 0 + y = 2 .
x +1
1
c) y 00 + 4y = .
cos 2t

Content

2.3 Euler equations
Inhomogeneous equations with constant coefficients
Form:
y 00 + py 0 + qy = f (x), p, q ∈ R.
Solve the corresponding homogeneous equation
y 00 + py 0 + qy = 0.
(We use the characteristic equation).

Find a particular solution to inhomogeneous equation.
General approach: variation of parameters.
Special RHS ⇒ special approach: undetermined coefficients
f (x) = e αx Pn (x) and
f (x) = e αx (Pn (x) cos βx + Qm (x) sin βx), where
Pn (x), Qm (x) are polynomials of x.

The case f (x) = e αx Pn (x)
Consider the RHS of the form
f (x) = e α x Pn (x),
where Pn (x) = an x n + an−1 x n−1 + . . . + a0 is a polynomial of

degree n.
Compare α to the root of the characteristic equation. We look for
a particular solution y ∗ to the inhomogeneous equation of the
form
if α is not a root: y ∗ = e αx Qn (x),
if α is a single root, y ∗ = xe αx Qn (x),
if α is a double root, y ∗ = x 2 e αx Qn (x),
where Qn (x) is a polynomial of degree n.
Plug y ∗ in the equation to find the coefficients of Qn (x).

When substituting y ∗ = e αx Q(x) into the equation, we obtain

q× y ∗ = e αx Q(x)
p× (y ) = αe αx Q(x) + e αx Q 0 (x)
∗ 0
(y ∗ )00 = α2 e αx Q(x) + 2αe αx Q 0 (x) + e αx Q 00 (x)

Hence, dividing both sides by e αx , we gain
(α2 + pα + q)Q(x) + (2α + p)Q 0 (x) + Q 00 (x) = Pn (x).
Example
Solve the following equations:
1 y 00 + 4y 0 + 5y = 8e −2x .
2 y 00 − 2y 0 = xe 2x + 3.
3 y 00 − 2y 0 + y = 2e x , y (0) = −1, y 0 (0) = 0.

Example
Solve the ODE: y 00 + 4y 0 + 5y = 8e −2x .
The characteristic equation is k 2 + 4k + 5 = 0

⇒ k1,2 = −2 ± i. The general solution to the corresponding
homogeneous equation is ȳ = e −2x (C1 cos x + C2 sin x).
α = −2 is not a solution of (CE). We look for y ∗ = Ae −2x .
Plugging in the equation, we have:
5× y∗ = Ae−2x
4× (y ∗ )0 = −2Ae −2x
(y ∗ )00 = 4Ae −2x
8e −2x = Ae −2x ⇒ A = 8; y ∗ = 8e −2x .
The general solution to the given ODE is
y = e −2x (C1 cos x + C2 sin x) + 8e −2x .

Example
Solve the ODE: y 00 − 2y 0 = xe 2x .
The characteristic equation is k 2 − 2k = 0 ⇒ k1 = 0, k2 = 2.

The general solution to the corresponding homogeneous
equation is ȳ = C1 + C2 e 2x .
α = 2 is a single root of (CE). We look for a particular
solution y ∗ = x(Ax + B)e 2x .
0× y∗ = (Ax 2 + Bx)e 2x
−2× (y ∗ )0 = 2(Ax 2 + Bx)e 2x + (2Ax + B)e 2x
(y ∗ )00 = 4(Ax 2 + Bx)e 2x + 2.(2Ax + B).2e 2x + 2Ae 2x
xe 2x = 2(2Ax + B)e 2x + 2Ae 2x
2
⇒ A = 41 , B = − 41 ; y ∗ = x 4−x e 2x .
2
y = C1 + C2 e 2x + x 4−x e 2x .
Example
Solve the IVP: y 00 − 2y 0 + y = 2e x , y (0) = −1, y 0 (0) = 0.
The characteristic equation is k 2 − 2k + 1 = 0 ⇒ k1,2 = 1.

The general solution to the corresponding homogeneous
equation is ȳ = (C1 + C2 x)e x .
α = 1 is a double root of (CE). We look for a particular
solution y ∗ = x 2 Ae x .
1× y∗ = Ax 2 e x
−2× (y ∗ )0 = Ax 2 e x + 2Axe x
(y ∗ )00 = Ax 2 e x + 2.2Ax.e x + 2Ae x
2e x = 2Ae x ⇒ A = 1; y ∗ = x 2 e x .
The general solution is y = (C1 + C2 x)e x + x 2 e x .
y 0 (x) = (C1 + C2 x)e x + C2 e x + x 2 e x + 2xe x .
Using the initial data −1 = y (0) = C1 ;
0 = y 0 (0) = C1 + C2 ⇒ C2 = 1.
Hence, the solution in search is y = (x 2 + x − 1)e x .

αx
The case f (x) = e Pm (x) cos βx + Qn (x) sin βx
Consider the RHS of the form

f (x) = e α x Pm (x) cos β x + Qn (x) sin β x ,
where Pm (x), Qn (x) are polynomials of degrees m, n.

Compare α ± iβ to the root of the characteristic equation. We
look for y ∗ of the form

if α ± iβ is not a root, y ∗ = e αx P̄l (x) cos βx + Q̄l (x) sin βx ,
if α ± iβ isa single root,
y ∗ = xe αx P̄l (x) cos βx + Q̄l (x) sin βx ,
where P̄l (x), Q̄l (x) are polynomials of degree l, l = max{m, n}.
Plug y ∗ in the equation to find the coefficients of P̄l (x), Q̄l (x).

Example
Solve the following equations:
1 y 00 + 2y 0 + 5y = 17 cos 2x.
2 y 00 + y = 2 cos x, y (0) = 2, y 0 (0) = −1.
3 y 00 + 4y = xe −2x − 3 sin 2x.

Example
Solve the ODE: y 00 + 2y 0 + 5y = 17 cos 2x.
The characteristic equation is k 2 + 2k + 5 = 0

⇒ k1,2 = −1 ± 2i. The general solution to the corresponding
homogeneous equation is ȳ = e −x (C1 cos 2x + C2 sin 2x).
α = 0, β = 2, ±2i is not a solution of (CE). We look for
y ∗ = A cos 2x + B sin 2x.
5× y∗ = A cos 2x + B sin 2x
∗
2× (y ) = 0 2B cos 2x − 2A sin 2x
(y ∗ )00 = −4A cos 2x − 4B sin 2x
( 2x = (A + 4B) cos 2x + (B − 4A) sin 2x
17 cos
A + 4B = 17
⇒ ⇒ A = 1, B = 4; y ∗ = cos 2x + 4 sin 2x.
B − 4A = 0
y = e −x (C1 cos 2x + C2 sin 2x) + cos 2x + 4 sin 2x.
Example
Solve the following IVP y 00 + y = 2 cos x, y (0) = 2, y 0 (0) = −1.
The characteristic equation is k 2 + 1 = 0 ⇒ k1,2 = ±i. The general

solution to the corresponding homogeneous equation is
ȳ = C1 cos x + C2 sin x.
α = 0, β = 1, ±i is a solution of (CE). We look for
y ∗ = x(A cos x + B sin x).
Plugging in the given equation
y∗ = x(A cos x + B sin x)
(y ∗ )00 = −x(A cos x + B sin x) + 2(−A sin x + B cos x)
⇒ 2 cos x = 2(−A sin x + B cos x) ⇒ A = 0, B = 1; y ∗ = x sin x.
The general solution to the inhomogeneous equation is
y = C1 cos x + C2 sin x + x sin x.
Using the initial data y (0) = 2 ⇒ C1 = 2.
y 0 = −C1 sin x + C2 cos x + sin x − x cos x, y 0 (0) = −1 ⇒ C2 = −1.
Hence, the solution in search is y = 2 cos x − sin x + x sin x.
Example
Solve the ODE y 00 + 4y = xe −2x − 3 sin 2x.
C.E. is k 2 + 4 = 0 ⇒ k1,2 = ±2i. Hence,

ȳ = C1 cos 2x + C2 sin 2x.
We look for y ∗ = y1∗ + y2∗ , where y1∗ and y2∗ are solutions of
the equations y 00 + 4y = xe −2x và y 00 + 4y = −3 sin 2x,
respectively.
α = −2 is not a solution of C.E., y1∗ = (Ax + B)e −2x .
α = 0, β = 2, 2i is a root of the C.E.,
y2∗ = x(A cos 2x + B sin 2x).
The general solution of the given equation is y = ȳ + y1∗ + y2∗ .

Plugging in the equations

y1∗ = (Ax + B)e −2x
(y1∗ )00 = 4(Ax + B)e −2x + 2.A.(−2)e −2x
⇒ xe −2x = 8(Ax + B)e −2x − 4Ae −2x
⇒ A = 18 , B = 16
1
; y1∗ = 2x+1
16 e
−2x .
∗
y2 = x(A cos 2x + B sin 2x)
∗ 00
(y2 ) = −4(A cos 2x + B sin 2x) + 2(−2A sin 2x + B cos 2x)
⇒ −3 sin 2x = −4A sin 2x + 4B cos 2x
⇒ A = 34 , B = 0; y2∗ = 43 x cos 2x.
The general solution is
y = C1 cos 2x + C2 sin 2x + 2x+1 −2x + 3 x cos 2x.
16 e 4

Content

2.3 Euler equations
Euler equations
Form
x 2 y 00 + axy 0 + by = f (x), a, b ∈ R.
Set x = e t , x > 0 ⇒ t = ln x.
1
y 0 (x) =y 0 (t).t 0 (x) = y 0 (t)
x
1 0 1 1 dy 0
y 00 (x) = y 0 (t) = − 2 y 0 (t) +
x x x x dt x
1 00 0
= 2 (y (t) − y (t)).
x
The given equation is transformed to
y 00 (t) + (a − 1)y 0 (t) + by (t) = g (t).
(Second order linear equation with constant coefficients).

Example
Solve the equations
1 x 2 y 00 − 9xy 0 + 21y = 36x ln x.
2 x 2 y 00 − 2xy 0 + 2y = 3x 2 .
Set ln x = t, then xy 0 (x) = y 0 (t); x 2 y 00 (x) = y 00 (t) − y 0 (t).

The equation becomes y 00 − 10y 0 + 21y = 36te t .
C.E. is k 2 − 10k + 21 = 0 ⇒ k1 = 3, k2 = 7; ȳ = C1 e 3t + C2 e 7t .
y ∗ = (At + B)e t , plugging in the equation, we obtain
A = 3, B = 2, y ∗ = (3t + 2)e t .
The general solution of the given equation is
y = C1 e 3t + C2 e 7t + (3t + 2)e t = C1 x 3 + C2 x 7 + x(3 ln x + 2).

Content

2.3 Euler equations
Mass – spring system: forced oscillation
Consider a vertical mass – spring system, the mass is acted by an

external force g (t). The equation governing the system is
mẍ(t) + γ ẋ(t) + kx(t) = g (t)
where mẍ - the force of inertia, γ ẋ - the damping force, kx -spring

force, g (t) - external force acting on the mass.
g (t): the input or driving force; x(t): the output or the response
of the system.
Undamped case: γ = 0.
Damped case: γ 6= 0.
Special interest: g (t) = F0 cos ωt, ω > 0.

Forced vibrations without damping
mẍ(t) + kx(t) = F0 cos ωt

We consider two interesting cases:
q
k
If ω0 = m 6= ω, then the general solution is
F0
x = C1 cos ω0 t + C2 sin ω0 t + cos ωt.
m(ω02 − ω2)
In particular, the mass is initially at rest, i.e. x(0) = 0,

x 0 (0) = 0. We obtain

F0 (cos ωt − cos ω0 t) 2F0 (ω0 − ω)t (ω0 + ω)t
x= = sin sin .
m(ω02 − ω 2 ) m(ω02 − ω 2 ) 2 2
If |ω0 − ω| is small, then the motion is a rapid oscillation with

high frequency, but a slowly varying sinusoidal amplitude.

If ω0 = ω, then the general solution is

F0
x = C1 cos ω0 t + C2 sin ω0 t + t sin ω0 t.
2mω0
Because of the term t sin ω0 t, it is predicted that the motion
will become unbounded regardless of the initial data.
Actually, as x is large, the model is invalid.

If damping is included in the model, the predicted motion remains

bounded; however, the response to the input function F0 cos ωt
may be quite large if γ is small and ω ≈ ω0 (resonance).
Resonance can be either good or bad depending on the
circumstances. It must be taken very seriously in the design of
structures, such as buildings or bridges, where it can produce
instabilities possibly leading to the catastrophic failure of the
structure. For example, soldiers traditionally break step when
crossing a bridge to eliminate the periodic force of their marching
that could resonate with a natural frequency of the bridge. On the
other hand, resonance can be put to good use in the design of
instruments, such as seismographs, intended to detect weak
periodic incoming signals.

Forced vibrations with damping
mẍ(t) + γ ẋ(t) + kx(t) = F0 cos ωt

Exer. 65 states that all solutions of the corresonding homogeneous
F0
equation approach 0. x(t) = C1 e r1 t + C2 e r2 t + cos(ωt − δ),
| {z } ∆
transient solution
| {z }
steady state
q 2 −ω)2
m(ω
where ∆ = m2 (ω02 − ω)2 + γ 2 ω 2 ; cos δ = 0
∆ ; sin δ = γω
∆.
With damping, the energy put into the system by the initial
displacement and velocity is dissipated, and the motion becomes
the response of the system to the external force.
Without damping, the effect of the initial data would persist for all
time.

Further investigation on the steady state solution
F0
X (t) = cos(ωt − δ)
∆
F0
The amplitude R = √ .
m2 (ω02 −ω)2 +γ 2 ω 2
For low frequency ω → 0, R → Fk0 .
For high frequency ω → ∞, R → 0.
m(ω 2 −ω)2
The phase angle δ: cos δ = 0
∆ ; sin δ = γω
∆.
For low frequency ω → 0, δ → 0, the response is nearly in
phase with the excitation.
For ω = ω0 , δ = π2 , the response lags behind the excitation by
π
2.
For high frequency ω → ∞, δ ≈ π, the response is nearly out
of phase with the excitation, i.e. the response attains its
minima when the excitation attains its maxima.

Tuan 11.12 InhomogeneouslinearODE

Uploaded by

Copyright:

Available Formats

Tuan 11.12 InhomogeneouslinearODE

Uploaded by

Document Information

Original Title

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

Tuan 11.12 InhomogeneouslinearODE

Uploaded by

Copyright:

Available Formats

Chapter 5: Second order linear ODEs (continued)

Nguyen Thu Huong

School of Applied Mathematics and Informatics

2. Inhomogeneous second order linear equations

2nd order linear DEs Nguyen Thu Huong

y 00 + p(x)y 0 + q(x)y = f1 (x) + f2 (x).

2nd order linear DEs Nguyen Thu Huong

y 00 + p(x)y 0 + q(x)y = f (x) (1)

has the form y = ȳ + y ∗ , where

Last week: Solve for ȳ = C1 y1 (x) + C2 y2 (x). In the special case of

2nd order linear DEs Nguyen Thu Huong

2. Inhomogeneous second order linear equations

ȳ (x) = C1 y1 (x) + C2 y2 (x).

We look for a particular solution of the inhomogeneous equation in

A particular solution of the inhomogeneous equation is found in

C10 (x)y10 (x) + C20 (x)y20 (x) = f (x).

2nd order linear DEs Nguyen Thu Huong

How to solve y 00 + p(x)y 0 + q(x)y = f (x)

1 Solve the homogeneous equation y 00 + p(x)y 0 + q(x)y = 0. If

2nd order linear DEs Nguyen Thu Huong

2nd order linear DEs Nguyen Thu Huong

2. Inhomogeneous second order linear equations

Inhomogeneous equations with constant coefficients

Solve the corresponding homogeneous equation

(We use the characteristic equation).

2nd order linear DEs Nguyen Thu Huong

The case f (x) = e αx Pn (x)

Consider the RHS of the form

where Pn (x) = an x n + an−1 x n−1 + . . . + a0 is a polynomial of

2nd order linear DEs Nguyen Thu Huong

When substituting y ∗ = e αx Q(x) into the equation, we obtain

(y ∗ )00 = α2 e αx Q(x) + 2αe αx Q 0 (x) + e αx Q 00 (x)

(α2 + pα + q)Q(x) + (2α + p)Q 0 (x) + Q 00 (x) = Pn (x).

2nd order linear DEs Nguyen Thu Huong

The characteristic equation is k 2 + 4k + 5 = 0

2nd order linear DEs Nguyen Thu Huong

The characteristic equation is k 2 − 2k = 0 ⇒ k1 = 0, k2 = 2.

The characteristic equation is k 2 − 2k + 1 = 0 ⇒ k1,2 = 1.

Consider the RHS of the form

where Pm (x), Qn (x) are polynomials of degrees m, n.

2nd order linear DEs Nguyen Thu Huong

2nd order linear DEs Nguyen Thu Huong

The characteristic equation is k 2 + 2k + 5 = 0

The characteristic equation is k 2 + 1 = 0 ⇒ k1,2 = ±i. The general

C.E. is k 2 + 4 = 0 ⇒ k1,2 = ±2i. Hence,

2nd order linear DEs Nguyen Thu Huong

Plugging in the equations

2nd order linear DEs Nguyen Thu Huong

2. Inhomogeneous second order linear equations

y 00 (t) + (a − 1)y 0 (t) + by (t) = g (t).

(Second order linear equation with constant coefficients).

Set ln x = t, then xy 0 (x) = y 0 (t); x 2 y 00 (x) = y 00 (t) − y 0 (t).

2nd order linear DEs Nguyen Thu Huong

2. Inhomogeneous second order linear equations

Mass – spring system: forced oscillation

Consider a vertical mass – spring system, the mass is acted by an

mẍ(t) + γ ẋ(t) + kx(t) = g (t)

where mẍ - the force of inertia, γ ẋ - the damping force, kx -spring

2nd order linear DEs Nguyen Thu Huong