Variable Coe Cient, Second Order, Linear, Ordinary Differential Equations

Variable Coecient, Second Order, Linear,
Ordinary Dierential Equations

Many physical, chemical and biological systems can be described using mathemat-
ical models. Once the model is formulated, we usually need to solve a dierential
equation in order to predict and quantify the features of the system being mod-
elled. As a precursor to this, we consider linear, second order ordinary dierential
equations of the form
P(x)
d
2
y
dx
2
+ Q(x)
dy
dx
+ R(x)y = F(x),
with P(x), Q(x) and R(x) nite polynomials that contain no common factor. This
equation is inhomogeneous and has variable coecients. The form of these poly-
nomials varies according to the underlying physical problem that we are studying.
However, we will postpone any discussion of the physical origin of such equations
until we have considered some classical mathematical models in Chapters 2 and 3.
After dividing through by P(x), we obtain the more convenient, equivalent form,
d
2
y
dx
2
+ a
1
(x)
dy
dx
+ a
0
(x)y = f(x). (1.1)
This process is mathematically legitimate, provided that P(x) = 0. If P(x
0
) = 0
at some point x = x
0
, it is not legitimate, and we call x
0
a singular point of the
equation. If P(x
0
) = 0, x
0
is a regular or ordinary point of the equation. If
P(x) = 0 for all points x in the interval where we want to solve the equation, we
say that the equation is nonsingular, or regular, on the interval.
We usually need to solve (1.1) subject to either initial conditions of the form
y(a) = , y
(a) = or boundary conditions, of which y(a) = and y(b) =

are typical examples. It is worth reminding ourselves that, given the ordinary dif-
ferential equation and initial conditions (an initial value problem), the objective
is to determine the solution for other values of x, typically, x > a, as illustrated in
Figure 1.1. As an example, consider a projectile. The initial conditions are the po-
sition of the projectile and the speed and angle to the horizontal at which it is red.
We then want to know the path of the projectile, given these initial conditions.
For initial value problems of this form, it is possible to show that:
(i) If a
1
(x), a
0
(x) and f(x) are continuous on some open interval I that contains
the initial point a, a unique solution of the initial value problem exists on
the interval I, as we shall demonstrate in Chapter 8.
4 VARIABLE COEFFICIENT, SECOND ORDER DIFFERENTIAL EQUATIONS
calculate the
solution for x > a
given y(a) =
and y'(a) =
y
a
x
Fig. 1.1. An initial value problem.
(ii) The structure of the solution of the initial value problem is of the form
y = Au
1
(x) +Bu
2
(x)

Complementary function
+ G(x)

Particular integral
,
where A, B are constants that are xed by the initial conditions and u
1
(x)
and u
2
(x) are linearly independent solutions of the corresponding homoge-
neous problem y
+a
1
(x)y
+a
0
(x)y = 0.
These results can be proved rigorously, but nonconstructively, by studying the
operator
Ly
d
2
y
dx
2
+a
1
(x)
dy
dx
+a
0
(x)y,
and regarding L : C
2
(I) C
0
(I) as a linear transformation from the space of
twice-dierentiable functions dened on the interval I to the space of continuous
functions dened on I. The solutions of the homogeneous equation are elements
of the null space of L. This subspace is completely determined once its basis
is known. The solution of the inhomogeneous problem, Ly = f, is then given
formally as y = L
1
f. Unfortunately, if we actually want to construct the solution
of a particular equation, there is a lot more work to do.
Before we try to construct the general solution of the inhomogeneous initial value
problem, we will outline a series of subproblems that are more tractable.
1.1 THE METHOD OF REDUCTION OF ORDER 5
1.1 The Method of Reduction of Order
As a rst simplication we discuss the solution of the homogeneous dierential
equation
d
2
y
dx
2
+ a
1
(x)
dy
dx
+ a
0
(x)y = 0, (1.2)
on the assumption that we know one solution, say y(x) = u
1
(x), and only need to
nd the second solution. We will look for a solution of the form y(x) = U(x)u
1
(x).
Dierentiating y(x) using the product rule gives
dy
dx
=
dU
dx
u
1
+ U
du
1
dx
,
d
2
y
dx
2
=
d
2
U
dx
2
u
1
+ 2
dU
dx
du
1
dx
+ U
d
2
u
1
dx
2
.
If we substitute these expressions into (1.2) we obtain
d
2
U
dx
2
u
1
+ 2
dU
dx
du
1
dx
+ U
d
2
u
1
dx
2
+ a
1
(x)
dU
dx
u
1
+ U
du
1
dx
+ a
0
(x)Uu
1
= 0.
We can now collect terms to get
U
d
2
u
1
dx
2
+ a
1
(x)
du
1
dx
+ a
0
(x)u
1
+ u
1
d
2
U
dx
2
+
dU
dx
2
du
1
dx
+ a
1
u
1
= 0.
Now, since u
1
(x) is a solution of (1.2), the term multiplying U is zero. We have
therefore obtained a dierential equation for dU/dx, and, by dening Z = dU/dx,
have
u
1
dZ
dx
+ Z
2
du
1
dx
+ a
1
u
1
= 0.
Dividing through by Zu
1
we have
1
Z
dZ
dx
+
2
u
1
du
1
dx
+ a
1
= 0,
which can be integrated directly to yield
log |Z| + 2 log |u
1
| +
x
a
1
(s) ds = C,
where s is a dummy variable, for some constant C. Thus
Z =
c
u
2
1
exp
x
a
1
(s) ds
=
dU
dx
where c = e
C
. This can then be integrated to give
U(x) =
x
c
u
2
1
(t)
exp
t
a
1
(s) ds
dt + c,
for some constant c. The solution is therefore
6 VARIABLE COEFFICIENT, SECOND ORDER DIFFERENTIAL EQUATIONS
y(x) = u
1
(x)
_
x
c
u
2
1
(t)
exp
_
_
t
a
1
(s) ds
_
dt + cu
1
(x).
We can recognize cu
1
(x) as the part of the complementary function that we knew
to start with, and
u
2
(x) = u
1
(x)
_
x
1
u
2
1
(t)
exp
_
_
t
a
1
(s)ds
_
dt (1.3)
as the second part of the complementary function. This result is called the reduc-
tion of order formula.
Example
Lets try to determine the full solution of the dierential equation
(1 x
2
)
d
2
y
dx
2
2x
dy
dx
+ 2y = 0,
given that y = u
1
(x) = x is a solution. We rstly write the equation in standard
form as
d
2
y
dx
2

2x
1 x
2
dy
dx
+
2
1 x
2
y = 0.
Comparing this with (1.2), we have a
1
(x) = 2x/(1 x
2
). After noting that
_
t
a
1
(s) ds =
_
t
2s
1 s
2
ds = log(1 t
2
),
the reduction of order formula gives
u
2
(x) = x
_
x
1
t
2
exp
_
log(1 t
2
)
_
dt = x
_
x
dt
t
2
(1 t
2
)
.
We can express the integrand in terms of its partial fractions as
1
t
2
(1 t
2
)
=
1
t
2
+
1
1 t
2
=
1
t
2
+
1
2(1 + t)
+
1
2(1 t)
.
This gives the second solution of (1.2) as
u
2
(x) = x
_
x
_
1
t
2
+
1
2(1 + t)
+
1
2(1 t)
_
dt
= x
_
1
t
+
1
2
log
_
1 + t
1 t
__
x
=
x
2
log
_
1 + x
1 x
_
1,
and hence the general solution is
y = Ax + B
_
x
2
log
_
1 + x
1 x
_
1
_
.

Variable Coe Cient, Second Order, Linear, Ordinary Differential Equations

Uploaded by

Copyright:

Available Formats

Variable Coe Cient, Second Order, Linear, Ordinary Differential Equations

Uploaded by

Document Information

Original Description:

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

Variable Coe Cient, Second Order, Linear, Ordinary Differential Equations

Uploaded by

Copyright:

Available Formats

Variable Coecient, Second Order, Linear,

Ordinary Dierential Equations

(a) = or boundary conditions, of which y(a) = and y(b) =

You might also like