MA 102 (Ordinary Differential Equations)

IIT Guwahati

Tutorial Sheet No. 9 Date: March 23, 2020

Applications of ODEs, Theory of higher order ODEs

1. Assuming that the rate of decay of a radioactive substance is proportional to the amount of sub-
stance present at any time t, show that in general the half life of a radioactive substance is

(t2 − t1 ) ln 2
t=
ln(A1 /A2 )

where A1 = A(t1 ) and A2 = A(t2 ), t1 < t2 .

Solution: Since a radioactive substance decays at a rate proportional to the amount of sub-
stance present at any time t, therefore the equation describing this process is
dA
= kA(t), (∗)
dt
where k is a constant of proportionality. Upon solving (∗) we have A(t) = A0 ekt , where we
assume that A(0) = A0 . Thus A1 = A(t1 ) = A0 ekt1 and A2 = A(t2 ) = A0 ekt2 . Hence
A2 k(t2 −t1 ) ⇒ k(t − t ) = ln( A2 ) ⇒ k = 1 A2
A1 = e 2 1 A1 t2 −t1 ln( A1 ). Let t0 be the half life of the
A0 ln(1/2) ln(2)(t2 −t1 )
substance. Therefore A0 ekt0 = 2 ⇒ t0 = k = ln(A1 /A2 ) .

2 2
2. Show that the length of the part of the ellipse xa2 + yb2 = 1 (a > b) that lies in the first quadrant is
R a q a2 −e2 x2
0 a2 −x2
dx where e is the eccentricity of the ellipse.

Solution: The length of the arc of the ellipse in the first quadrant is given by
Z a
dy
S0 = [1 + ( )2 ]1/2 dx.
0 dx
x2 y2 2y dy 2x dy 2 2
From a2
+ b2
= 1,we have b2 dx
+ a2
=0⇒ = − ab 2 xy = − ab 2 q x x2 = − a√abx
dx 2 −x2
. Thus
b 1− 2
q a
Ra 2 2 R a 4 2 −a2 )x2 1/2 Ra
S0 = 0 [1 + a2 (ab 2x−x2 ) ]1/2 dx = 0 [ a a+(b
2
2 (a2 −x2 ) ] dx = 0 {a2 − (1 − ab 2 )x2 }/(a2 − x2 )dx =
R a q a2 −e2 x2
0 a2 −x2
dx.

3. Light strikes a plane curve C in such a manner that all beams L parallel to the y-axis are reflected
to a single point O. Determine the shape of the curve.
Let θ be the angle of incidence at any point P (x, y) of the curve (refer to the figure), of a ray AP
parallel to the y-axis. Then the tangent at P makes an angle ( π2 − θ) with the positive direction
of the x-axis. Also let φ be the angle made by the reflected ray P O with the y-axis. Then from
the triangle OBP , π − φ + 2θ = π ⇒ φ = 2θ. Also, tanθ = cot( π2 − θ) = tan( 1π −θ) = dy 1
= dx
dy and
2 dx
y
tan( π2 − φ) = x ⇒ tanφ = x
y. Thus φ = 2θ ⇒ tanφ = 2tanθ
1−tan2 θ
and

x 2 dx
dy
⇒ =  2
y
1 − dxdy
 2
x x dx dx
⇒ − =2
y y dy dy
 2
x dx dx x
⇒ +2 − =0
y dy dy y
 2
dx dx
⇒ x + 2y =x (1)
dy dy

Thus the curve C is the solution of (1). Let x2 = w. Then 2x dx

dy =
dw
dy and (1) takes the form

x dw 2 y dw
( ) + =x
4x2 dy x dy
 2
dw dw
⇒ + 4y + 4y 2 = 4(w + y 2 )
dy dy
 2
dw
⇒ + 2y = 4(w + y 2 )
dy
dw p
⇒ + 2y = ±2 w + y 2 .
dy
p dv dw
Let v = w + y 2 . Then 2v dy = dy + 2y and the equation takes the form

dv
2v = ±2v ⇒ dv = ±dy ⇒ v = ±y + c.
dy
p
Therefore, w + y 2 = c ± y ⇒ x2 = c2 ± 2cy.
4. If y = φ1 (x) is a particular solution of y 00 + (sinx)y 0 + 2y = ex and y = φ2 (x) is a particular solution
of y 00 + (sinx)y 0 + 2y = cos(2x), then find a particular solution of y 00 + (sinx)y 0 + 2y = ex + 2sin2 x.

Solution: φ1 (x) − φ2 (x) + 12 .

5. Determine the largest interval (a, b) in which the given IVP is certain to have a unique solution:
y0
(a) ex y 00 − x−3 + 3y = ln x, y(1) = 3, y 0 (1) = 2.
(b) (1 − x)y − 3xy 0 + 3y = sinx, y(0) = 1, y 0 (0) = 1.
00

(c) x2 y 00 + 4y = cosx, y(1) = 0, y 0 (1) = −1.

Solution: (a) (0, 3); (b) (−∞, 1); (c) (0, ∞).

6. If f = (f1 , f2 , . . . , fn ) and u = (u1 , u2 , . . . , un ) be two n-tuples of functions fi , ui ∈ C n (R), 1 ≤ i ≤ n,

such that Af = u for some n × n real matrix A, then show that

W (u1 , u2 , . . . , un )(x) = det A × W (f1 , f2 , . . . , fn )(x).

Use the above to answer the following.

(a) If u = 3g + f and v = −g − f, then W (u, v) = 4(tcost − sint) where W (f, g) = tcost − sint.
(b) If y1 and y2 are linearly independent solutions of y 00 + p(t)y 0 + q(t)y = 0, under what conditions
will y3 = a1 y1 + a2 y2 and y4 = b1 y1 + b2 y2 also be linearly independent solutions?

Solution: Since Af = u, assuming that A = [aij ]n×n , we have ui = Σnj=1 aij fj , i = 1, 2, . . . , n.

(k) (k)
Thus ui = Σnj=1 aij fj , i ∈ {1, 2, . . . , n}, k ∈ {0, 1, . . . , n − 1} so that

 Σnj=1 a1j fj Σnj=1 a2j fj ··· Σnj=1 anj fj

 

 Σnj=1 a1j fj0 Σnj=1 a2j fj0 Σnj=1 anj fj0
 
··· 
W (u1 , . . . , un ) =  .. .. ..
 

 . . ··· . 
 Σn a f (n−1) Σn a f (n−1) (n−1)
 
j=1 1j j j=1 2j j ··· Σnj=1 anj fj 

(n−1)
 
 Σn a1j fj
 j=1 Σnj=1 a1j fj0 ··· Σnj=1 a1j fj 

(n−1)
Σnj=1 a2j fj0
 n
 Σj=1 a2j fj ··· Σnj=1 a2j fj


=  .. .. ..

···

 . . . 
(n−1)
 
 Σn anj fj
j=1 Σnj=1 anj fj0 ··· Σnj=1 anj fj 
  (n−1)

f10

 a11 a12 ··· a1n  f1 ··· f1 

(n−1)
 
 a21 a22 ··· a2n f20 ···

 f2 f2 
= .
 
.. .. .. .. .. 
 ..

. ··· . . . ··· .
 
  
 an1 an2 ··· ann (n−1) 
fn fn0 ··· fn
 

= det(A)W (f1 , f2 , . . . , fn )
 
 3 1 
(a) W (u, v) =  W (g, f ) = 4(tcost − sint).
−11 
 
a1 a2 

(b) W (y3 , y4 )(t) =  W (y1 , y2 )(t) and W (y1 , y2 )(t) 6= 0 in any interval where p(t) and
b1 b2 
 
 a1 a2 
q(t) are continuous. Since y1 , y2 are linearly independent solutions, so are y3 , y4 if   6= 0.
b1 b2 
 7. Let p(x), q(x) ∈ C(I). Assume that the functions y1 , y2 ∈ C 2 (I) are solutions of the differential
equations y 00 + p(x)y 0 + q(x)y = 0 on an open interval I. Prove that if y1 and y2 are zero at the
same point in I, then they cannot be a fundamental set of solutions on that interval.

Solution: Since y1 (x0 ) = y2 (x0 ) = 0 for some x0 ∈ I, we find that W (y1 , y2 )(x0 ) = 0 for some
Rx
x0 ∈ I. But we have from Abel’s formula that W (y1 , y2 )(x) = W (y1 , y2 )(x0 )exp(− x0 p(t)dt)
for all x ∈ I. Hence, W (y1 , y2 )(x) = 0 ∀x ∈ I ⇒ y1 and y2 cannot form a fundamental set of
solutions on I.

000
8. Let S = {f : R → R |L(f ) = 0}, where L(f ) := f + f 00 − 2. Find the Ker(L). Let S0 ⊂ Ker(L)
be the subspace of solutions g such that limx→∞ g(x) = 0. Find g ∈ S0 such that g(0) = 0 and
g 0 (0) = 2.

Solution: The auxiliary equation (AE) r3 + r2 − 2 = 0 ⇒ (r − 1)(r2 + 2r + 2) = 0. Ker(L) =

span {ex , e−x cosx, e−x sinx}. f (x) ∈ Ker(L) has the form f (x) = c1 ex + c2 e−x cosx + c3 e−x sinx.
S0 is obtained by putting c1 = 0. Thus, g(x) ∈ S0 has the form g(x) = c2 e−x cosx + c3 e−x sinx.
Using the IC g(0) = 0 and g 0 (0) = 2, we obtain c2 = 0 and c3 = 2. So, g(x) = 2e−x sinx.

9. Consider the differential equation x2 y 00 − 4xy 0 + 6y = 0 on (−∞, ∞);

(a) Verify that the functions y1 (x) = x3 and y2 (x) = x2 |x| are linearly independent solutions.
(b) Show that y1 and y2 are linearly dependent on (−∞, 0), but are linearly independent on
(−∞, ∞);
(c) Although y1 and y2 are linearly independent, show that W (y1 , y2 ) = 0 for all x ∈ (−∞, ∞).
Does this violate the fact that W (y1 , y2 ) = 0 for every x ∈ (−∞, ∞) implies y1 and y2 are
linearly dependent?

Solution: y1 and y2 satisfy the given differential equation. On (−∞, 0), y2 = (−1)y1 , hence
they are linearly dependent. On (−∞, ∞), consider c1 x3 +c2 x2 |x| = 0. If x = 1 then c1 +c2 = 0,
and if x = −1, c1 − c2 = 0. This implies c1 = c2 = 0. Hence, y1 and y2 are linearly independent
on (−∞, ∞). Note that on 0 < x < ∞, W (y1 , y2 )(x) = 0. Thus, W (y1 , y2 ) = 0 on (−∞, ∞).
It doesn’t violate the fact. Observe that p(x) = −4/x and q(x) = 6/x2 fail to be continuous at
x = 0. Thus, the continuity assumption on the coefficients p and q cannot be dropped.

10. If y1 and y2 are linearly independent solutions of (x2 +1)y 00 +2xy 0 +xex y = 0 and if W (y1 , y2 )(1) = 2,
find the value of W (y1 , y2 )(5).

R 
5 2x 2
Solution: By Abel’s formula, W (y1 , y2 )(5) = W (y1 , y2 )(1) exp 1 − 1+x2 dx = 13 .