In this assignment, we will denote: Solutions of Assignment - 3

y �� + p(x)y � + q(x)y = r(x), x ∈ I (∗)

y �� + p(x)y � + q(x)y = 0, x ∈ I (∗∗)

where I ⊂ R is an interval and p(x), q(x), r(x) are continuous functions on I.

1. (T) Let y1 be the solution of the IVP

y �� + (2x − 1)y � + sin(ex )y = 0, y(0) = 1, y � (0) = −1;

and y2 be the solution of the IVP

y �� + (2x − 1)y � + sin(ex )y = 0, y(0) = 2, y � (0) = −1.

Find the Wronskian of y1 , y2 . What is the general solution of y �� +(2x−1)y � +sin(ex )y = 0?

Solution:
We know that if y1 , y2 are solutions of (∗∗), then the Wronskian W (y1 , y2 )(x) = W (x) =
� 2
c exp(− p(x)dx) = ce−x +x . From the given initial conditions we have W (0) = 3. So
2
c = 3. Hence W (x) = 3e−x +x .
Since W (0) �= 0, we deduce thaty1 , y2 are independent solutions. Therefore, the general
solution is given by c1 y1 + c2 y2 .

2. (T) Show that the set of solutions of the linear homogeneous equation (∗∗) is a real vector
space. Also show that the set of solutions of the linear non-homogeneous equation (∗)
is not a real vector space. If y1 (x), y2 (x) are any two solutions of (∗), obtain conditions
on the constants a and b so that ay1 + by2 is also its solution.
Solution:
Let S be the set of solutions of the linear homogeneous ODE (**). Clearly S is a subset
of set of twice differentiable functions on I which is a real vector space. Thus it is
sufficient to show that S is subspace of the above vector space of twice differentiable
function. Now 0(x) = 0 satisfies (**) and hence 0 ∈ S. Thus S is nonempty. Also if u, v
both satisfies (**), then αu(x) + v(x) is also a solution of (**). This implies αu + v ∈ S.
Hence S is a subspace, i.e. a vector space.
Now 0(x) = 0 is not a solution of (*), thus zero element does not exist. Hence, the set
of solution of (*) is not a real vector space.
Let y1 (x), y2 (x) are any two solutions of (∗). Then

y1�� + p(x)y1� + q(x)y1 = r(x), (1)

y2�� + p(x)y2� + q(x)y2 = r(x). (2)
 Multiplying (1) by a and (2) by b and adding, we find

(ay1 + by2 )�� + p(x)(ay1 + by2 )� + q(x)(ay1 + by2 ) = (a + b)r(x).

If ay1 + by2 is also a solution, then the LHS is r(x) and hence a + b = 1.

3. Decide if the statements are true or false. If the statement is true, prove it, if it is false,
give a counter example showing it is false.
(i) If f (x) and g(x) are linearly independent functions on an interval I, then they are
linearly independent on any larger interval containing I.
If f (x) and g(x) are linearly independent functions on an interval I, then they are linearly
independent on any smaller interval contained in I.
(ii) If f (x) and g(x) are linearly dependent functions on an interval I, then they are
linearly dependent on any subinterval of I.
If y1 (x) and y2 (x) are linearly dependent functions on an interval I, then they are linearly
dependent on any larger interval containing I.
(iii) If y1 (x) and y2 (x) are linearly independent solution of (∗∗) on an interval I, they
are linearly independent on any interval contained in I.
(iv) If y1 (x) and y2 (x) are linearly dependent solutions of (∗∗) on an interval I, they are
linearly dependent on any interval contained in I.
Solution:
(i) True, follows from the definition of linear independence. Flase: take f (x) = x2 and
g(x) = x|x|. Then f, g linearly independent over [−1, 1] but dependent over [0, 1].
(ii)True, follows from definition.
(iii) True, follows from the fact that, in this case y1 , y2 is linearly independent on I iff
W (y1 , y2 ) �= 0 on all I.
(iv) True, follows from the fact that, in this case y1 , y2 is linearly dependent on I iff
W (y1 , y2 ) = 0 on all I.

4. Can x3 be a solution of (∗∗) on I = [−1, 1]? Find two 2nd order linear homogeneous
ODE with x3 as a solution.
Solution:
No. Putting y = x3 in the given equation, we get 6x + p(x)3x2 + q(x)x3 = 0 for all
x ∈ [−1, 1.] Cancelling x, we get 6 + p(x)3x + q(x)x2 = 0 for all [−1, 1] � x �= 0. That is
p(x)3 + q(x)x = −6/x for all x ∈ [−1, 1.]. We see that LHS is continuous at 0 but RHS
is not continuous at 0. This cant not happen.
Two ODEs with x3 as solution are: xy �� = 2y � and x2 y �� = 6y. Note that here p, q are
not continuous at 0.
5. (T) Can x sin x be a solution of a second order linear homogeneous equation with con-
stant coefficients?
Solution:
No, putting x sin x in y �� + py � + qy = 0, we get (q − 1)x sin x + p(sin x + x cos x) = 0 for
all x ∈ R. Here p, q are constants. This is clearly not possible.

6. (T) Find the largest interval on which a unique solution is guaranteed to exist of the
IVP. (x + 2)y �� + xy � + cot(x)y = x2 + 1, y(2) = 11, y � (2) = −2.
Solution:
Comparing with (∗), we have

x cos(x) x2 + 1
p(x) = , q(x) = , r(x) = .
x+2 (x + 2) sin x x+2
The discontinuities of p, q, r are x = −2, 0, ±π, ±2π, ±3π, · · · . The largest interval that
contains x0 = 2 but none of the discontinuities is, therefore, (0, π).

7. Without solving determine the largest interval in which the solution is guaranteed to
uniquely exist of the IVP ty �� − y � = t2 + t, y(1) = 1, y � (1) = 5. Verify your answer by
solving it explicitly.
Solution:
Since p, r are not continuous at 0, the maximum interval of existence and uniqueness of
solution of the given IVP is (0, ∞).
Here dependent variable y is missing. Solving it, y(t) = t3 /3 + 7t2 /4 + t2 (ln t)/2 − 13/12
for which the max interval of validity is (0, ∞).

8. Find the differential equation satisfied by each of the following two-parameter families
of plane curves:
(i) y = cos(ax + b) (ii) y = ax + xb (iii) y = aex + bxex
Solution:
For two arbitrary constants, the order of the ODE will be two. Eliminate constants a
and b by differentiating twice.
(i) y = cos(ax + b) =⇒ y � = −a sin(ax + b), y �� = −a2 cos(ax + b) = −a2 y. From this
we find
y �2 �2 2 y
��
+ y 2
= 1 =⇒ (1 − y 2 2
)a = y =⇒ −(1 − y ) = y �2 =⇒ (1 − y 2 )y �� + yy �2 = 0
a2 y

(ii) y = ax + b/x =⇒ xy = ax2 + b =⇒ xy � + y = 2ax =⇒ y � + y/x = 2a which on

differentiating again gives y �� + y � /x − y/x2 = 0 =⇒ x2 y �� + xy � − y = 0.
(iii) y = aex + bxex =⇒ e−x y = a + bx =⇒ e−x y � − e−x y = b =⇒ e−x y �� − 2e−x y � +
e−x y = 0 which on simplification gives y �� − 2y � + y = 0
 9. Find general solution of the following differential equations given a known solution y1 :
(i) (T) x(1 − x)y �� + 2(1 − 2x)y � − 2y = 0 y1 = 1/x
2 �� �
(ii) (1 − x )y − 2xy + 2y = 0 y1 = x
Solution:
(i) Here y1 = 1/x. Substitute y = u(x)/x to get (1 − x)u�� − 2u� = 0. Thus, u� =
1/(1 − x)2 and u = 1/(1 − x). Hence, y2 = 1/(x(1 − x)) and the general solution is
y = a/x + b/(x(1 − x)).
(ii) Here y1 = x. Substitute y = xu(x) to get x(1 − x2 )u�� = 2(2x2 − 1)u� . Thus,

u�� 2(2x2 − 1) 2 1 1 1
�
= 2
=− − + =⇒ u� = 2
u x(1 − x ) x 1+x 1−x x (1 − x2 )

Thus, � � � �
� 1 1 1 1 1 1 1+x
u = 2+ + =⇒ u = − + ln(
x 2 1+x 1−x x 2 1−x
Hence, � �
x 1+x
y2 = −1 + ln(
2 1−x
and the general solution is
� � ��
x 1+x
y = ax + b −1 + ln( .
2 1−x
√
10. Verify that sin x/ x is a solution of x2 y �� + xy � + (x2 − 1/4)y = 0 over any interval on
the positive x-axis and hence find its general solution.
Solution:
Verification is straightforward.
√
Substitute y = u(x) sin x/ x to get
� �
sin x �
� cos x sin x
y = √ u + √ − 3/2 u
x x 2x
� � � �
�� sin x �� cos x sin x � sin x cos x 3 sin x
y = √ u + 2 √ − 3/2 u + − √ − 3/2 + u
x x 2x x x 4 x5/2
This leads to

sin x u�� + 2 cos x u� = 0 =⇒ u� = cosec2 x =⇒ u = − cot x

√ √
Hence, y2 = − cos x/ x and the general solution is y = (a sin x + b cos x)/ x.

11. Solve the following differential equations:

(i) y �� − 4y � + 3y = 0 (ii) y �� + 2y � + (ω 2 + 1)y = 0, ω is real.
Solution:
 (i) Characteristic (or auxiliary) equation: m2 − 4m + 3 = 0 =⇒ m = 1, 3.
General sol: y = Aex + Be3x
(ii) Characteristic equation: m2 + 2m + (1 + ω 2 ) = 0 =⇒ m = −1 ± ωi.
Case 1: ω = 0 =⇒ equal roots m = −1, −1 and general sol: y = (A + Bx)e−x
Case 2: ω �= 0 =⇒ complex conjugate roots m = −1 ± ωi and general sol: y =
e−x (A sin ωx + B cos ωx)

12. Solve the following initial value problems:

(i) (T) y �� + 4y � + 4y = 0 y(0) = 1, y � (0) = −1
(ii) y �� − 2y � − 3y = 0 y(0) = 1, y � (0) = 3
Solution:
(i) Assume y = emx is a solution. Putting in the given equation, we get the characteristic
equation: m2 + 4m + 4 = 0 =⇒ m = −2, −2. General sol: y = e−2x (A + Bx). Using
initial conditions:
A = 1, B − 2A = −1 =⇒ B = 1 =⇒ y = (x + 1)e−2x

(ii) Characteristic equation: m2 − 2m − 3 = 0 =⇒ m = −1, 3. General sol: y =

(Ae3x + Be−x ). Using initial conditions:
A + B = 1, 3A − B = 3 =⇒ A = 1, B = 0 =⇒ y = e3x

13. Reduce the following second order differential equation to first order differential equation
and hence solve.
(i) xy �� + y � = y �2 (ii) (T) yy �� + y �2 + 1 = 0 (iii) y �� − 2y � coth x = 0
Solution:
(i) Dependent variable y absent. Substitute y � = p =⇒ y �� = dp/dx. Thus xp� + p = p2 .
Solving p = 1/(1 − ax) which on integrating again gives y = b − ln(1 − ax)/a, where a
and b are arbitrary constants.
(ii) Independent variable x is absent in yy �� + y �2 + 1 = 0. Substitute y � = p =⇒ y �� =
p dp/dy. Thus
dp pdp dy � a2
py + p2 = 1 =⇒ + = 0 =⇒ ln 1 + p2 y = ln a =⇒ 1 + p2 =
dy 1 + p2 y y2
From p2 = a2 /y 2 − 1, we find
ydy �
� = ±dx =⇒ − a2 − y 2 = ±x + b.
a2 − y 2
Both the solutions can be written as (x + b)2 + y 2 = a2 where a and b are arbitrary
constants..
(iii) y �� − 2y � coth x = 0. Substitute y � = p =⇒ y �� = dp/dx. Thus dp/dx = 2p coth x.
Solving p = a sinh2 x, which on integrating again gives y = a(sinh 2x − 2x)/4 + b where
a and b are arbitrary constants.
14. Find the curve y = y(x) which satisfies the ODE y �� = y � and the line y = x is tangent
at the origin.
Solution:
The given conditions lead to the following problem:
Solve y �� − y � = 0 with y(0) = 0, y � (0) = 1. Integrating once gives y � − y = a which on
another integration gives y + a = bex . y(0) = 0 gives a = b. y � (0) = 1 gives b = 1 and
hence solution is y = ex − 1.

15. Are the following functions linearly dependent on the given intervals?
(i) sin 4x, cos 4x (−∞, ∞) (ii) ln x, ln x3 (0, ∞)
(iii) cos 2x, sin2 x (0, ∞) (iv)(T) x3 , x2 |x| [−1, 1]
Solution:
(i) a sin 4x + b cos 4x = 0. For x = 0 we find b = 0 and for x = π/8 we get a = 0. Hence
they are NOT linearly dependent.
(ii) ln x3 − 3 ln x = 0 for x ∈ (0, ∞). Hence linearly dependent.
(iii) a cos 2x + b sin2 x = 0. For x = 0 we find a = 0 and for x = π/2 we get b = 0. Hence
they are NOT linearly dependent.
(iv) ax3 + bx2 |x| = 0. For x < 0 we find a − b = 0 and for x > 0 we get a + b = 0. Hence
a = b = 0 and thus they are NOT linearly dependent.

16. (a) Show that a solution to (**) with x-axis as tangent at any point in I must be
identically zero on I.
(b) (T) Let y1 (x), y2 (x) be two solutions of (**) with a common zero at any point in I.
Show that y1 , y2 are linearly dependent on I.
(c) (T) Show that y = x and y = sin x are not a pair solutions of equation (**), where
p(x), q(x) are continuous functions on I = (−∞, ∞).
Solution:
(a) Let ξ(x) be the solution. Since x axis is a tangent, at x = x0 , say, then ξ(x0 ) =
ξ � (x0 ) = 0. Clearly y(x) ≡ 0 satisfies (**) and the initial conditions y(x0 ) = y � (x0 ) = 0.
Since the solution is unique, ξ(x) ≡ 0 in I .
(b) If y1 (x), y2 (x) have a common zero at x = x0 , say, then y1 (x0 ) = y2 (x0 ) = 0. Hence,
W (y1 , y2 ) = 0 at x = x0 and thus y1 , y2 are linearly dependent.
(c) y1 = x and y2 = sin x are LI on I. So if they were solution of (∗∗), the wronskian
W (y1 , y2 ) must never be zero. But W (y1 , y2 ) = 0 at x = 0, a contradiction.

17. (a)(T) Let y1 (x), y2 (x) be two twice continuously differentiable functions on an interval
I.
(i) Show that the Wronskian W (y1 , y2 ) does not vanish anywhere in I if and only if there
exists continuous p(x), q(x) on I such that (**) has y1 , y2 as independent solutions.
 (ii) Is it true that if y1 , y2 are independent on I then there exists continuous p(x), q(x)
on I such that (**) has y1 , y2 as independent solutions?
(b) Construct equations of the form (**) from the following pairs of solutions: e−x , xe−x .
Solution:
(a)(i) Suppose that W (y1 , y2 ) does not vanish anywhere in I. We want to find p(x), q(x)
such that

y1�� + p(x)y1� + q(x)y1 = 0, y2�� + p(x)y2� + q(x)y2 = 0. (3)

Solving we get:
d
p(x) = −(y1 y2�� − y2 y1�� )/W (y1 , y2 ) = − (W (y1 , y2 ))/W (y1 , y2 )
dx
and q(x) = (y1� y2�� − y2� y1�� )/W (y1 , y2 ). They are continuous on I since W (y1 , y2 ) never
zero on I.
� �
[Note that q(x) can also be written as q(x) = − y11 y1�� + p(x)y1� .]
Converse follows from the fact Wronskian is never zero for independent solutions of (∗∗).
(ii) Not true. Consider y1 (x) = x3 and y2 (x) = x2 |x| on I = [−1, 1.] Then they are
independent on I, but they are not solutions of any (∗∗) on I.
(b) Using 8(a): y1 (x) = e−x and y2 (x) = xe−x . Hence, W (y1 , y2 ) = e−2x and p(x) = 2.
And q(x) = −(e−x − 2e−x )/e−x = 1. Hence y �� + 2y � + y = 0.
Alternative: Write y = ay1 (x) + by2 (x) and eliminate a and b. y = e−x (a + bx) =⇒
ex y = a + bx. Differentiating w.r.t. x twice we find

ex (y � + y) = b =⇒ ex (y �� + 2y � + y) = 0 =⇒ y �� + 2y � + y = 0

18. By using the method of variation of parameters, find the general solution of:
(i) y �� + 4y = 2 cos2 x + 10ex (ii) (T) y �� + y = x sin x
(iii) y �� + y = cot2 x (iv) x2 y �� − x(x + 2)y � + (x + 2)y = x3 , x > 0.
[Hint. y = x is a solution of the homogeneous part]
Solution:
If y1 , y2 are independent solutions of the homogeneous part of the ODE

y �� + p(x)y � + q(x)y = r(x),

then the general solution is y = Ay1 +By2 +uy1 +vy2 , where A, B are arbitrary constants
and � �
ry2 ry1
u=− dx, v = dx, [W (y1 , y2 ) is the Wronskian]
W W
(i) y1 = cos 2x, y2 = sin 2x, W (y1 , y2 ) = 2, r(x) = 2 cos2 x + 10ex = cos 2x + 1 + 10ex .
Now �
cos 4x cos 2x
u = − y2 r/W dx = + − ex (sin 2x − 2 cos 2x)
16 4
 �
sin 4x x sin 2x
v= y1 r/W dx = + + + ex (2 sin 2x + cos 2x)
16 4 4
Thus
cos 2x x sin 2x 1
yp = + + + 2ex
16 4 4
General solution: (absorbing first term of yp in the homogeneous solution)
x sin 2x 1
y = A cos 2x + B sin 2x + + + 2ex
4 4

(ii) y1 = cos x, y2 = sin x, W (y1 , y2 ) = 1, r(x) = x sin x. Now

�
x2 x sin 2x cos 2x
u = − y2 r/W dx = − + +
4 4 8
�
x cos 2x sin 2x
v = y1 r/W dx = − +
4 8
Thus
cos x x sin x x2 cos x
yp = + −
8 4 4
General solution: (absorbing first term of yp in the homogeneous solution)

x sin x x2 cos x
y = A cos x + B sin x + −
4 4
(iii) (ii) y1 = cos x, y2 = sin x, W (y1 , y2 ) = 1, r(x) = cot2 x. Now
�
u = − y2 r/W dx = − ln(cosecx − cot x) − cos x
�
v= y1 r/W dx = −cosecx − sin x

Thus
yp = −2 − cos x ln(cosecx − cot x)
General solution:

y = A cos x + B sin x − 2 − cos x ln(cosecx − cot x)

(iv) y1 = x is a solution of the homogeneous part. To find another linearly independent

solution we assume y = xu. This gives

u�� − u� = 0 =⇒ u� − u = 1 =⇒ u = ex − 1 =⇒ y = xex − x

Since y1 = x, we take y2 = xex . The nonhomogeneous part is written as

x + 2 � (x + 2)
y �� − y + y = x.
x x2
Thus r(x) = x and W (y1 , y2 ) = x2 ex . Now
�
u = − y2 r/W dx = −x
 and �
v= y1 r/W dx = −e−x

Thus yp = −x − x2 .
General solution: (absorbing first term of yp in the homogeneous solution)

y = x(A + Bex ) − x2 .

19. Find the general solution of a 7th-order homogeneous linear differential equation with
constant coefficients whose characteristic polynomial is p(m) = m(m2 − 3)2 (m2 + m + 2).
Solution:
√ √ √
m = 0, ± 3, ± 3, −1/2 ± i 7/2. So general solution:
√ √ √ √ √ √
y = c 1 + c2 e 3x
+ c3 xe 3x
+ c4 e − 3x
+ c5 e − 3x
+ c6 e−x/2 cos( 7x/2) + c7 e−x/2 sin( 7x/2).

Initial Value Problem vs. Boundary Value Problem

A second-order initial value problem consists of a second-order ordinary differential equation
y �� (t) = F (t, y(t), y � (t)) and initial conditions y(t0 ) = y0 , y � (t0 ) = y0� where t0 , y0 , y0� are
numbers.
It might seem that there are more than one ways to present the initial conditions of a
second order equation. Instead of locating both initial conditions y(t0 ) = y0 and y � (t0 ) = y0�
at the same point t0 , couldn’t we take them at different points, for examples y(t0 ) = y0 and
y(t1 ) = y1 ; or y � (t0 ) = y0� and y � (t1 ) = y1� ? The answer is NO. All the initial conditions in
an initial value problem must be taken at the same point t0 . The sets of conditions
above where the values are taken at different points are known as boundary conditions. A
boundary value problem does not have the existence and uniqueness guaranteed.
Example: Every function of the form y = C sin(t), where C is a real number satisfies the
boundary value problem y �� + y = 0, y(0) = 0, y(π) = 0. Therefore, the problem has infinitely
many solutions, even though p(t) = 0, q(t) = 1, r(t) = 0 are all continuous everywhere.