ODE: Assignment-5 4 solutions

(For calculations of Particular Integrals by operator method, see Simmons books, page 161,
section 23 of the chapter Second order linear equations.)

1. Solve: (i) x2 y �� + 2xy � − 12y = 0 (ii)(T) x2 y �� + 5xy � + 13y = 0 (iii) x2 y �� − xy � + y = 0

d2 y dy
[Recall: The ODE of the form x2 2
+ ax + by = 0, where a, b are constants, is
dx dx
called the Cauchy-Euler equation. Under the transformation x = et (when x > 0) for
d2 y dy
the independent variable, the above reduces to 2 + (a − 1) + by = 0, which is an
dt dt
equation with constant coefficients. ]
Solution:
(i) Using the substitution x = et , the given equation reduces to

d2 u du
2
+ −12u = 0 =⇒ m2 +m−12 = 0 =⇒ m = −4, 3 =⇒ u(t) = Ae−4t +Be3t = y(et ).
dt dt
The general solution is thus
A
y(x) = + Bx3 .
x4
(ii) Using the substitution x = et , the given equation reduces to,

d2 u du
2
+ 4 + 13u = 0 =⇒ m2 + 4m + 13 = 0 =⇒ m = −2 ± 3i.
dt dt
Thus
u(t) = e−2t (A cos 3t + B sin 3t) = y(et ).
The general solution is
1� �
y(x) = 2
A cos(3 ln x) + B sin(3 ln x) .
x

(iii) Using the substitution x = et , the given equation reduces to

d2 u du
2
− 2 + u = 0 =⇒ m2 − 2m + 1 = 0 =⇒ m = 1, 1 =⇒ u(t) = et (A + Bt) = y(et )
dt dt
The general solution is thus
y(x) = ex (A + B ln x).

d
2. (Higher order Cauchy-Euler equations) Let us denote D = dx and D = dtd where x = et .
Show that
xD = D, x2 D2 = D(D − 1), x3 D3 = D(D − 1)(D − 2).
Hence conclude that (x3 D + ax2 D2 + bxD + c)y = 0, x > 0 is transformed into constant
coefficients ODE [D(D − 1)(D − 2) + aD(D − 1) + bD + c]y = 0 by the substitution
x = et .
 Solution:
Given x = et , so dx
dt
d
= et = x. Now, by chain rule, dx = dtd dx
dt
= e−t dtd . Thus xD =
dt
D. Differentiating this with respect to x, we have xD2 + D = D2 dx = D2 e−t , =⇒
x2 D2 + xD = D2 , =⇒ x2 D2 = D2 − D = D(D − 1).
Differentiating x2 D2 = D2 − D with respect to x, we have x2 D3 + 2xD2 = [D3 − D2 ]e−t ,
=⇒ x3 D3 = D3 − D2 − 2D(D − 1) = D(D − 1)(D − 2).

3. Find a particular solution of each of the following equations by operator methods and
hence find its general solution:
(i) y �� + 4y = 2 cos2 x + 10ex (ii)(T) y �� + y = sin x + (1 + x2 )ex
(T) (iii) y �� − y = e−x (sin x + cos x) (iv) y ��� − 3y �� − y � + 3y = x2 ex
Solution:
(i) Characteristic equation m2 + 4 = 0 =⇒ m = ±2i. Hence homogeneous solution
yh = A cos 2x + B sin 2x. Now r(x) = 2 cos2 x + 10ex = cos 2x + 1 + 10ex . Let D ≡ d/dx
and yp be the particular solution. Then

1 1
2
1 = 2
e0x = 1/4.
D +4 D +4
1 1
10ex = 10 2 ex = 2ex .
D2+4 1 +4
1 1 2ix 1 2ix
2
e2ix = x e =x e = xe2ix /4i.
D +4 2D 2.2i
Taking real part
1
cos 2x = x sin 2x/4.
D2 +4
Adding, we get the particular solution as
x sin 2x 1
yp = + + 2ex .
4 4

Thus the general solution is

x sin 2x 1
y = A cos 2x + B sin 2x + + + 2ex .
4 4

(ii) Characteristic equation m2 + 1 = 0 =⇒ m = ±i. Hence homogeneous solution

yh = A cos x + B sin x. Now r(x) = sin x + (1 + x2 )ex . Let D ≡ d/dx and yp be the
particular solution. Then
1 1 ix 1 x
2
eix = x e = x eix = (cos x + i sin x).
1+D 2D 2i 2i
Taking imaginary part
1 x cos x
sin x = − .
1 + D2 2
 1 1 1
2
(1 + x2 )ex = ex 2
(1 + x2 ) = ex 2 (x2 + 1)
1+D (D + 1) + 1 D + 2D + 2

ex 1 2 ex
= (x + 1) = (1 − D − D2 /2 + (D + D2 /2)2 + · · · )(x2 + 1)
2 1 + D + D2 /2 2

ex ex ex
= (1−D−D2 /2+D2 /2+· · · )(x2 +1) = (1−D+D2 /2+· · · )(x2 +1) = (1+x2 −2x+1)
2 2 2
Thus the general solution is
� �
x cos x x2 x
y = A cos x + B sin x − + 1−x+ e
2 2

(iii) Characteristic equation m2 − 1 = 0 =⇒ m = ±1. Hence homogeneous solution

yh = Aex + Be−x . Now r(x) = e−x (sin x + cos x). Let D ≡ d/dx and yp be the particular
solution.

1 −x ix e−x eix e−x eix e−x

e e = = = − [cos x + 2 sin x + i(sin x − 2 cos x)].
D2 − 1 (i − 1)2 − 1 −2i − 1 5

Then the particular solution is obtained by adding the real and imaginary parts:
e−x (cos x − 3 sin x)
yp (x) =
5
Thus the general solution is
e−x (cos x − 3 sin x)
y = Aex + Be−x +
5
(iv) Characteristic equation m3 − 3m2 − m + 3 = 0 =⇒ m = −1, 1, 3. Hence homoge-
neous solution yh = Ae−x + Bex + Ce3x . Now r(x) = x2 ex . Let D ≡ d/dx and yp be the
particular solution. Then

1 1 1
x2 e x = x2 e x = e x 3 x2
D3 2
− 3D − D + 3 3
(D − 1) − 4(D − 1) D − 4D
1 x 1 x 1 1 e x x3 x
= ex x 2
= e (1 + D 2
/4 + · · · )x 2
= e (x 2
+ ) = − ( + ).
−4D(1 − D2 /4) −4D −4D 2 4 3 2
So the particular integral is
� �
x x x3
yp (x) = −e + .
8 12
Thus the general solution is
� �
−x x 3x x x x3
y = Ae + Be + Ce −e + .
8 12
4. Solve y �� + y � − 2y = ex .
Solution: Characteristic equation of the homogeneous part is: m2 + m − 2 = 0, m =
1, −2. Solution for the homogeneous part: c1 ex + c2 e−2x .
Particular integral:
1 x 1 x xex xex
e = x e = = .
D2 + D − 2 2D + 1 (2.1 + 1) 3

General solution:
x −2x xex
c 1 e + c2 e + .
3
5. Solve by using operator method (D2 + 9)y = sin 2x cos x.
Solution:
Characteristic equation of the homogeneous part is: m2 + 9 = 0, m = ±3i. Solution for
the homogeneous part: c1 cos 3x + c2 sin 3x.
Particular integral:
1 1
sin 2x cos x = (sin 3x + sin x).
D2 +9 2
2(D + 9)

Now
1
eix = eix /(i2 + 9) = (cos x + i sin x)/8.
+9D2
Taking the imaginary part,
1
(sin x) = sin x/16.
2(D2 + 9)

Now
1 3ix xe3ix
e = = (x cos 3x + ix sin 3x)/6i.
D2 + 9 2.3i
Taking the imaginary part,
1
(sin 3x) = −(x cos 3x)/12.
2(D2 + 9)

General solution:

c1 cos 3x + c2 sin 3x + sin x/16 − (x cos 3x)/12.

6. Find a particular integral by operator method: D2 − 6D + 9 = 1 + x + x2 .

Solution:

1 1
P.I = 1 + x + x2 = 1 + x + x2
D2 − 6D + 9 2
9(1 + (D − 6D)/9)
 1
= [1 − (D2 − 6D)/9 + (D2 − 6D)2 /81 − · · · ](1 + x + x2 ).
9
1
= [1 + 2D/3 + D2 /3 + · · · ](1 + x + x2 ) = (1 + x + x2 + 2/3 + 4x/3 + 2/3)
9
1
= (7/3 + 7x/3 + x2 ).
9
7. Find P.I: y �� + 9y = x cos x.
Solution:
Consider
1 1 1 1
xeix = eix x = eix 2 x = eix x
D2 +9 2
(D + i) + 9 D + 2iD + 8 2
8(1 + D /8 + iD/4)

1 1
= eix (1 − D2 /8 − iD/4)x = eix (x − i/4).
8 8
Taking the real part:
1 x cos x sin x
x cos x = + .
D2 +9 8 32
8. (T) Solve x2 y �� − 2xy � − 4y = x2 + 2 log x, x > 0.
Solution:
Apply the transformation x = et the equation reduces to y �� − 3y − 4 = e2t + 2t.
Solution of the homogeneous part c1 e4t + c2 e−t .
1 1
Particular integral: D 2 −3D−4
(e2t + 2t) = −e2t /6 + D 2 −3D−4
2t = − 16 e2t − 12 (t − 3/4).

1 1 1 −1
2t = 2 t = − [1+(D2 /4−3D/4+· · · )]t = (t−3/4).
D2 − 3D − 4 2
−4(1 − D /4 + 3D/4) 2 2

Hence the general solution is:

1 1 1 1
y = c1 e4t + c2 e−t − e2t − (t − 3/4) = c1 x4 + c2 /x − x2 − (ln x − 3/4).
6 2 6 2

9. (T) (Higher order variation of parameter) Consider the n-th order linear equation
n
�
(n)
y + ai (x)y (i) = y (n) + an−1 (x)y (n−1) + · · · + a0 (x)y = r(x).
1

Assume that y1 , · · · , yn are n-independent solutions of the associated homogeneous equa-

tion. Prove that a particular integral of the given ODE is
� Ri
yp = vi yi where vi� = .
W
Here W is the wronskian of y1 , · · · , yn and Ri is the determinant obtained by replacing
i-th column of W by [0, 0, · · · , 0, r(x)].
 Solution:
Let
�
yp = vi yi − − − − − −(1).
� � � � �
Differentiating yp� = vi yi + vi yi� . Assume vi yi = 0 then
�
yp� = vi yi� − − − − − (2).
� � � �
Differentiating this yp�� = vi� yi� + vi yi�� . Assuming vi yi = 0 we have
�
yp�� = vi yi�� . − − − − − −(3)

Proceeding similarly, we get

� (n−1)
yp(n−1) = vi yi − − − −(n)
� (n−2)
if vi� yi = 0.

� (n−1)
� (n)
yp(n) = vi� yi + vi yi − − − − − −(n + 1).

Then n
� � (n−1)
yp(n) + ai (x)yp(i) = vi� yi .
1

Hence yp is a solution of the given ODE if

� � � � (n−2) � (n−1)
vi� yi = 0, vi� yi� = 0, vi� yi�� = 0, · · · , vi� yi = 0, vi� yi = r(x).

Solution such system of linear equation is given by vi� = R W

i
where W is the wronskian
of y1 , · · · , yn and Ri is the determinant obtained by replacing i-th column of W by
[0, 0, · · · , 0, r(x)].

10. (i) Let y1 (x), y2 (x) are two linearly independent solutions of y �� +p(x)y � +q(x)y = 0. Show
that φ(x) = αy1 (x) + βy2 (x) and ψ(x) = γy1 (x) + δy2 (x) are two linearly independent
solutions if and only if αδ �= βγ.
(ii) Show that the zeros of the functions a sin x + b cos x and c sin x + d cos x are distinct
and occur alternately whenever ad − bc �= 0.
Solution:
(i) We have W (φ, ψ) = (αδ − βγ)W (y1 , y2 ). Since y1 , y2 are fundamental solutions,
W (y1 , y2 ) �= 0. If αδ �= βγ, then W (φ, ψ) �= 0. Conversely if W (φ, ψ) �= 0, then
αδ �= βγ.
(ii) We know sin x, cos x are independent solutions of y �� + y = 0. So by part (i) a sin x +
b cos x and c sin x + d cos x are independent solutions whenever ad − bc �= 0. Hence the
result follows from Sturm Separation theorem ( Simmons, page 190, Theorem A).
11. (T) Show that any nontrivial solution u(x) of u�� + q(x)u = 0, q(x) < 0 for all x, has at
most one zero.
Solution:
Consider the equation z �� = 0. Then z = 1 is a solution of the equation. By Strum
comparison theorem, between two zeros of u(x) there must be at least one zero of z(x).
But z = 1 has no zero. Hence u(x) can have at most one zero.

12. Let u(x) be any nontrivial solution of u�� + [1 + q(x)]u = 0, where q(x) > 0. Show that
u(x) has infinitely many zeros.
Solution:
Consider
� �
v �� + v = 0, u�� + 1 + q(x) u = 0
Now v = sin x is a nontrivial solution of v �� + v = 0. Since 1 + q(x) > 1, by Strum
comparison theorem, u must vanish between two zeros of sin x. Since, sin x has infinitely
many zeros, u also has infinitely may zeros.

13. Let u(x) be any nontrivial solution of u�� + q(x)u = 0 on a closed interval [a, b]. Show
that u(x) has at most a finite number of zeros in [a, b].
Solution:
Suppose, on the contrary, u(x) has infinite number of zeros in [a, b]. It follows that there
exists x0 ∈ [a, b] and a sequence of zeros xn �= x0 such that xn → x0 . Since u(x) is
continuous and differentiable at x0 , we have

u(xn ) − u(x0 )
u(x0 ) = lim u(xn ) = 0, u� (x0 ) = lim =0
xn →x0 xn →x0 xn − x0
By uniqueness theorem, u ≡ 0 which contradicts the fact that u is nontrivial.

14. (T) Let Jp be any non-trivial solution of the Bessel equation

x2 y �� + xy � + (x2 − p2 )y = 0, x > 0.

Show that Jp has infinitely many positive zeros.

Solution:
The normal form of Bessel equation is

1/4 − p2
u�� + (1 + )u = 0.
x2
2
Given p ≥ 0, we can choose x0 large enough such that 1+ 1/4−px2
> 1/4 for all x ∈ (x0 , ∞).
�� 1
Compare Jp with sin(x/2) which is solution of v + 4 v = 0 in (x0 , ∞). Clearly sin(x/2) has
infinitely many zeros in (x0 , ∞). By Sturm comparison theorem, between two consecutive
zeros of sin(x/2) there is a zero of Jp . Hence Jp has infinitely many zero in (x0 , ∞).
15. (T) Consider u�� + q(x)u = 0 on an interval I = (0, ∞) with q(x) ≥ m2 for all t ∈ I.
Show any non trivial solution u(x) has infinitely many zeros and distance between two
consecutive zeros is at most π/m.
Solution: Compare u(x) with sin mx which is a solution of v �� + m2 v = 0. By Sturm
comparison theorem, between two consecutive zeros of v(x) = sin(mx) there is a zero of
u(x). Hence u(x) has infinitely many zero in (x0 , ∞).
Let u(a) = 0. We will show that u(x) has a zero in (a, a + π/m]. Consider v(x) =
sin(mx − ma) which is a solution of v �� + m2 v = 0. Clearly v(a) = v(a + π/m) = 0. Hence
by Sturm comparison theorem, there exists at least one zero of u(x) in (a, a + π/m).
Hence distance between two consecutive zeros of u(x) is at most π/m.

16. Consider u�� + q(x)u = 0 on an interval I = (0, ∞) with q(x) ≤ m2 for all t ∈ I. Show
that distance between two consecutive zeros is at least π/m.
Solution:
Suppose u(a) = 0 and u(b) be two consecutive zeros. Consider v(x) = sin(mx − ma)
which is a solution of v �� + m2 v = 0. By Sturm comparison theorem, there exists a zero
of v(x) in (a, b). But we know that v(a) = 0 and next zero of v is at a + π/m. So
b > a + π/m.

17. (T) Let Jp be any non-trivial solution of the Bessel equation

x2 y �� + xy � + (x2 − p2 )y = 0, x > 0.

Show that (i) If 0 ≤ p ≤ 1/2, then every interval of length π has at least contains at
least one zero of Jp .
(ii) If p = 1/2 then distance between consecutive zeros of Jp is exactly π.
(iii) If p > 1/2 then every interval of length π contains at most one zero of Jp .
Solution: The normal form of Bessel equation is

1/4 − p2
u�� + (1 + )u = 0.
x2
The zeros of Jp and u(x) are same.
(i)Apply exercise 15 with m = 1.
(ii) Clear from normal form.
(iii) Apply exercise 16 with m = 1.

18. Let y(x) be a non-trivial solution of y �� + q(x)y = 0. Prove that if q(x) > k/x2 for some
k > 1/4 then y has infinitely many positive zeros. If q(x) < 4x12 then y has only finitely
many positive zeros.
Solution:
Consider the Cauchy-Euler equation y �� + ky x2
= 0. With x = et , it transforms into
y �� − y � + ky = 0. So characteristic equation m2 − m + k = 0. So 1 − 4k = 0 implies two
equal real roots and so the solution has finitely many zeros. If 1 − 4k < 0 then complex
conjugate roots and solution look like xm sin(βx) and it has infinitely many zeros. Rest
follows from Sturm comparison theorem.