Professional Documents
Culture Documents
Solution Methods: Laplace Transform and Series Solutions of Odes
Solution Methods: Laplace Transform and Series Solutions of Odes
Solution Methods: Laplace Transform and Series Solutions of Odes
49
50 CHAPTER 3. SOLUTION METHODS
3. f (t) = tn , n ≥ 1.
4. f (t) = sin at, f (t) = cos at.
Theorem 3.1 (Existence and Uniqueness of Laplace Transform). If f (t)
is piecewise continuous and |f (t)| ≤ M ekt , ∀t ≥ 0, and some constants M and k
the Laplace transform L{f (t)} exists for all s < k.
1. F (s) = 3
s2 +4
. Soln: L−1 { s23+4 } = L−1 { 32 · 2
s2 +22
} = 23 L−1 { s2 +2
2 3
2 } = 2 sin 2t
2. F (s) = s+1
s2 +4
. Soln: L−1 { ss+1 1
2 +4 } = cos 2t + 2 sin 2t.
No. f (t) F (s) = L{f (t)} s No. f (t) F (s) = L{f (t)} s
1 1
1. 1 s
s > 0 2. t s2
s>0
n n! 1
3. t sn+1
s>0 4. eat s−a
s>0
n at n! a
5. t e (s−a)n+1
s>0 6. sin at s2 +a2
s>0
s 2as
7. cos at s2 +a2
s>0 8. t sin at (s2 +a2 )2
s>0
s2 −a2 s
9. t cos at (s2 +a2 )2
s>0 10. cosh at s2 −a2
s > |a|
a b
11. sinh at s2 −a2
s > |a| 12. eat sin bt (s−a)2 +b2
s>0
at s−a
13. e cos bt (s−a)2 +b2
s>0
Example 3.4. Show that L{eat sin bt} = (s−a)b2 +b2 and L{eat cos bt} = s−a
(s−a)2 +b2
.
b b
Solution: L{sin bt} = s2 +b 2 ⇒ F (s − a) = (s−a)2 +b2 .
Exercises 3.1.
First Derivative: If L{f (t)} = F (s), then L{f 0 (t)} = sF (s)−f (0) = sL{f (t)}−
f (0).
R∞ R∞
Proof. L{f 0 (t)} = 0
e−st f 0 (t)dt = e−st f (t)|∞
0 + 0
e−st f (t)dt = sL{f (t)} −
f (0).
Second Derivative: If L{f (t)} = F (s), then L{f 00 (t)} = s2 F (s)−sf (0)−f 0 (0) =
s2 L{f (t)} − sf (0) − f 0 (0).
Proof. As above L{f 00 (t)} = sL{f 0 (t)} − f 0 (0) = s2 L{f (t)} − sf (0) − f 0 (0).
Higher Order Derivative: Let L{f (t)} = F (s), then L{f n (t)} = sn F (s) −
sn−1 f (0) − sn−2 f 0 (0) − . . . − sf n−2 (0) − f n−1 (0).
Property IV: Let L{f (t)} = F (s), then L{−tf (t)} = F 0 (s). This also implies
n
that L{tf (t)} = −F 0 (s). In general, L{tn f (t)} = (−1)n d dsFn(s) .
R∞ R∞ R∞
Proof. F 0 (s) = ∂
∂s 0
e−st f (t)dt = 0
∂
∂s
(e−st )f (t)dt = 0
−te−st f (t)dt =
L{−tf (t)}.
52 CHAPTER 3. SOLUTION METHODS
Solution:
Step 1: Take Laplace transform on both sides: Let L{y(t)} = Y (s), and then
2 s2 − s + 2 s2 − s + 2
(s2 − 3s − 10)Y (s) = +s+2−3= , ⇒ Y (s) = .
s s s(s − 5)(s + 2)
Step 3: Take inverse Laplace transform to get y(t) = L−1 {Y (s)}. The main
technique here is partial fraction.
Structure of solutions:
• Take Laplace transform on both sides. You will get an algebraic equation for
Y (s).
2. y 00 −2y 0 +2y = e−t , y(0) = 0, y 0 (0) = 1. Ans. y(t) = 15 e−t − 15 et cos t+ 95 et sin t
3. y 00 + y = cos 2t, y(0) = 2, y 0 (0) = 1 Ans. y(t) = 73 cos t + sin t − 13 cos 2t.
Table 3.2: The terms in the partial fraction and their corresponding inverse Laplace
transform.
∞
X ∞
X
n−1
0
f (x) = ncn (x − a) = (n + 1)cn+1 (x − a)n .
n=1 n=0
Definition 3.5. If at a point x = a the functions p1 (x) and p2 (x) are analytic,
then the point a is said to be an ordinary point of the differential equation (3.1).
Further, if at x = a the functions p1 (x) and/or p2 (x) are not analytic, then a is
said to be a singular point of (3.1).
Definition 3.6. A singular point a at which the functions p(x) = (x − a)p1 (x) and
q(x) = (x − a)2 p2 (x) are analytic is called a regular singular point of the DE (3.1).
p(x) 0 q(x)
y 00 + y + y = 0, (3.2)
(x − a) (x − a)2
Example 3.7. If in the DE (3.1), p1 (x) and p2 (x) are constants, then every point
is an ordinary point.
x2 y 00 + a1 xy 0 + a2 y = 0, (3.3)
that is, the functions P = B/A and Q = C/A are analytic at x = a. Then (3.5)
has two linearly independent solutions, each of the form
∞
X
y(x) = an (x − a)n . (3.6)
n=0
The radius of convergence of any such series solution is at least as large as the
distance from a to the nearest (real or complex) singular point of (3.5). The
coefficients in the series in (3.6) can be determined by its substitution in (3.5).
Example 3.10. Determined the radius of convergence guaranteed by the above
theorem of the series solution of
(x2 + 9)y 00 + xy 0 + x2 y = 0.
the only singular points are x = ±3i. The distance of each of these from 0 is 3,
∞
an xn has the radius of convergence at least 3.
P
so a series solution of the form
n=0
The distance of each singular point from 4 is 5, so a series solution of the form
∞
an (x − 4)n , has radius of convergence at least 5.
P
n=0
We can begin the first and the third summations at n = 0 as well, because no
nonzero terms are thereby introduced. By shifting the index of the summation in
the second sum we get
∞
X ∞
X ∞
X ∞
X
n(n − 1)an xn − 4 (n + 2)(n + 1)an+2 xn + 3 nan xn + an xn = 0.
n=0 n=0 n=0 n=0
∞
X
n2 + 2n + 1 an − 4 (n + 2) (n + 1) an+2 xn = 0
n=0
(n + 1)2 an − 4 (n + 2) (n + 1) an+2 = 0,
d2 y dy
(t2 − 2t − 3) + 3(t − 1) + y = 0; y(1) = 4, y 0 (1) = 1. (3.12)
dt2 dt
∞
an (t − 1)n . But instead
P
Solution: We need a general solution of the form
n=0
of substituting this series in (3.12) to determine the coefficients, it simplifies the
computations if we first make the substitution x = t − 1, so our series solution looks
∞
an xn . To transform Eq. (3.12) into one with the new independent
P
of the form
n=0
variable x, we note that
t2 − 2t − 3 = (x + 1)2 − 2(x + 1) − 3 = x2 − 4
dy dy dx dy
= = = y0
dt dx dt dx
and
d2 y
d dx dx d 0
2
= ( ) = (y ) = y 00
dt dx dt dt dx
here the primes denotes differentiability with respect x. Hence Eq. (3.12) is
transformed into
Hence
1 1 1
y2 (x) = x − x3 − x4 + x5 + · · ·
6 12 30
Therefore, the Solution of the DE is
y(x) = y1 (x) + y2 (x)
Exercises 3.4.
2. Solve the initial value problems in the following problems, by first making
∞
an tn of
P
a substitution of the form t = x − a, and then find the solution
n
the transformed DE. State the interval of the values of x for which the above
theorem is guarantees convergence.
(a) y 00 + (x − 1)y 0 + y = 0; y(1) = 2, y 0 (1) = 0
(b) (2x − x2 )y 00 − 6(x − 1)y 0 − 4y = 0; y(1) = 0, y 0 (1) = 1
(c) (x2 − 6x + 10)y 00 − 4(x − 3)y 0 + 6y = 0; y(3) = 2, y 0 (3) = 0
3. In the following problems, find a three term recurrence relation for solutions
∞
an xn . Then find the first three nonzero terms in each of two
P
of the form
n
linearly independent solutions.
In this case we can verify by direct substitution that the simplest power function
y(x) = xr is a solution of (3.13) if and only if r is the root of the quadratic equation
r(r − 1) + p0 r + q0 = 0. (3.14)
In general case, in which p(x) and q(x) are power series rather than constants, it is
a reasonable conjecture that our differential equation might have a solution of the
form ∞ ∞
X X
r n
y(x) = x an x = an xn+r . (3.15)
n=0 n=0
assume that a0 =6 0, because the series must have a first nonzero term. Differenti-
ating (3.19) leads to
X∞
y0 = (n + r)an xn+r−1 . (3.20)
n=0
and ∞
X
00
y = (n + r)(n + r − 1)an xn+r−2 . (3.21)
n=0
Substituting of the series in (3.17) – (3.21) into (3.16) gives
[r(r − 1)a0 xr + (r + 1)rc1 xr+1 + · · · ]
+ [p0 x + P1 x2 + · · · ][ra0 xr−1 + (r + 1)a1 xr + · · · ]
+ [q0 + q1 x + q2 x2 + · · · ][a0 xr + a1 xr+1 + · · · ] = 0.
(3.22)
2
GEORG FERDINAND FROBENIUS (1849–1917) A German mathematician whose main
research was in group theory and analysis. He worked in Zurich and Berlin and published his
method for the series solution of linear ordinary differential equations in 1873.
3.2. SERIES SOLUTION OF ODES 63
r(r − 1) + p0 r + q0 = 0. (3.23)
Equation (3.23) is called the indicial equation of the differential in (3.17) and
its two roots (possibly equal) are the exponents of the differential equation (at
the regular singular point x = 0.)
Our derivation of (3.23) shows that if the Frobenius series is to be a solution of
the differential equation in (3.17), then the exponent r must be one of the roots
r1 and r2 the indicial equation in (3.23). If r1 6= r2 it follows that there are two
possible Frobenius series solutions, where as if r1 = r2 there is only one possible
Frobenius series solution.
(a) Roots r1 and r2 are real and distinct and do not differ by an integer
(b) Roots r1 and r2 are real and differ by an integer
(c) Roots r1 and r2 are real and equal
(d) Roots r1 and r2 are complex conjugates
Theorem 3.3 (Frobenius Series Solution). Suppose x = 0 is a regular singular
point of the equation
x2 y 00 + xp(x)y 0 + q(x)y = 0. (3.24)
Let ρ > 0 denote the minimum of radii of convergence of the power series,
∞
X ∞
X
p(x) = p n xn and q(x) = p n xn .
n=0 n=0
2. If r1 −r2 is neither zero nor positive integer, then there exists a second linearly
independent solution for x > 0 of the form
∞
X
y2 (x) = xr2 bn x n (b0 6= 0) (3.26)
n=0
The radii of convergence of the power series in (3.25) and (3.26) are each at least
ρ. The coefficients in these series can be determined by substituting the series in
the DE (3.24).
We have already seen that if r1 = r2 , then there can exists only one Frobenius
series solution it turns out that, if r1 − r2 is a positive integer there may or may not
exist a second Frobenius series solution of the second form in (3.26) corresponding
to the smaller root r2 . These exceptional cases will be discussed later.
Solution: First divide each term by the coefficient of y 00 to write the DE in the
form of
3
− 1 − 1 x2
y 00 + 2 y 0 + 2 2 2 y = 0.
x x
Now we see that x = 0 is a regular point. Since p(x) = 32 and q(x) = − 12 − 21 x2
then p0 = 32 and q0 = − 12 . Frobenius series we obtain will converge for all x > 0.
The indicial equation become
3 1 1
r(r − 1) + r − = (r + 1)(r − ) = 0,
2 2 2
so the roots of the indicial equation are r1 = 12 and r2 = −1. r1 − r2 = 3/2 is not
positive integer, so by the above theorem, there exist two linearly independent
Frobenius series solutions. These are of the form
∞
X ∞
X
n+1/2
y1 (x) = an x and y2 (x) = bn xn−1 ,
n=0 n=0
the series y1 , whereas with r2 = −1 it becomes a recurrence relation for the series
y2 . When substitute
∞
X ∞
X ∞
X
n+r 0 n+r−1 00
y= an x , y = (n+r)an x , and y = (n+r)(n+r −1)an xn+r−2
n=0 n=0 n=0
The case r1 = 12 ; Substituting r = 1/2 into the recurrence relation (3.29), we get
an−2
an = , for n ≥ 2.
2n2 + 3n
With this formula we can determine the coefficients in the first Frobenius solution
y1 . We see that an = 0 whenever n is odd. With n = 2, 4, · · · the formula for an
yields
a0 a2 a0 a4 a0
a2 = , a 4 = = , a6 = = .
14 44 616 90 55440
66 CHAPTER 3. SOLUTION METHODS
Hence
x2 x4 x6
y1 (x) = a0 x1/2 (1 + + + + · · · ).
14 616 55440
The case r2 = −1: We now replace bn in place of an and substitute r = −1 in
(3.29). This gives the recurrence relation
bn−2
bn = , for n ≥ 2.
2n2 − 3n
Again bn = 0 for n odd. With n = 2, 4, 6, · · · we get
b0 b0 a0
b2 = , b4 = , b6 =
2 40 2160
Hence the second Frobenius solution is
x2 x4 x6
y2 (x) = b0 x−1 (1 + + + + · · · ).
2 40 2160
Example 3.15. Find Frobenius series solutions of Bessel’s equation of order zero
x2 y 00 + xy 0 + x2 y = 0. (3.30)
Solution: First divide each term by the coefficient of y 00 , we get
1 x2
y 00 + y + 2 y = 0.
x x
Hence x = 0 is a regular singular point with p(x) = 1 and q(x) = x2 , so our series
will converges for all x > 0. Because p0 = 1 and q0 = 0, the indicial equation
becomes
r(r − 1) + r = r2 = 0.
Thus we obtain only the single exponent r = 0, and so there is only one Frobenius
series solution of the form ∞
X
y(x) = an x n
n=0
P∞
of the DE; which is a power series. Thus we substitute y(x) = n=0 an xn into the
DE the result is
∞
X ∞
X ∞
X
n n
n(n − 1)an x + nan x + an xn+2 = 0.
n=0 n=0 n=0
We collect the first two sums and shift the index of summation in the third by −2
to obtain ∞ ∞
X X
2 n
n an x + an−2 xn = 0.
n=0 n=0
3.2. SERIES SOLUTION OF ODES 67
−an−2
an = and n ≥ 2. (3.31)
n2
Because a1 = 0, we see that an = 0 whenever n odd. The substituting n = 2, 4, 6
into (3.31), we get
a0 a0 a0
a2 = − 2
, a4 = 2 2 and a6 = − 2 2 2 .
2 24 246
Evidently, the pattern is
(−1)n (−1)n a0
a2n = = .
22 .42 · · · (2n)2 22n (n!)2
The choice a0 = 1 gives us one of the most important special functions in mathe-
matics, the Bessel function of order zero of the first kind, denoted by J0 (x).
Thus
∞
X (−1)n x2n
J0 (x) = 2n (n!)2
.
n=0
2
xy 00 + 2y 0 + xy = 0. (3.32)
Reduction of Order
When only one Frobenius series solution exists, we need another technique. One
method is reduction of order. Consider the DE
p(x) 0 q(x)
y 00 + y + 2 y = 0, (3.35)
x x
is stated in the theorem below.
and
∞
X
r1
y2 (x) = y1 (x) ln x + x bn x n . (3.37)
n=1
∞
X
r1
y1 (x) = x an x n (a0 6= 0) (3.38)
n=0
and
∞
X
r2
y2 (x) = Cy1 (x) ln x + x bn x n . (3.39)
n=0
In (3.39) b0 6= 0 but C may be either zero or nonzero, so the logarithmic term may
or may not be present in this case. The radii of convergence of this theorem is
at least ρ. The coefficient in these series (and the constant C in (3.39)) may be
determined by direct substitution of the series in the differential equation (3.35).
x2 y 00 + xy 0 + x2 y = 0. (3.40)
and
∞
2y 0 y1 X
y200 = y100 ln x + 1 − 2 + n(n − 1)bn xn−2 .
x x n=2
We substitute these into the DE and use the fact that J0 (x) also satisfies this
equation to obtain
∞
X ∞
X ∞
X
0= (x2 y100 + xy10 +x 2
y1 ) ln x + 2xy10 + n(n − 1)bn x + n
nbn x n−2
+ bn xn+2
n=2 n=1 n=0
Now we examine the coefficients with even subscript in the above equation. First
we see that
(−1)2 1
b2 = −2 2 2 2
= .
2 2 (1!) 4
For n ≥ 2, we read the recurrence relation
2(−1)n (2n)
(2n)2 b2n + b2n−2 = − (3.43)
22n (n!)2
x2 y 00 − xy 0 + 10y = 0. (3.45)
Clearly, x=0 is a regular singular point with p(x) = −1 and q(x) = 10 and
p0 = −1, q0 = 10. So the indicial equation becomes
r(r − 1) − r + 10 = r2 − 2r + 10 = 0,
with complex conjugate solutions r = 1 ± 3i. Substituting
∞
X
y(x) = an xn+r ,
n=0
After terms are collected under a single summation sign, this becomes
∞
X
[(n + r)(n + r − 2) + 10]an xn+r = 0.
n=0
Exercises 3.5. Find the solutions (use the reduction of order when necessary) of
4. xy 00 + y 0 − xy = 0 8. x2 y − 4xy + 20y = 0