MAT306X/201/2008

DEPARTMENT OF MATHEMATICAL
SCIENCES

MATHEMATICS (MAT306X)
SOLUTIONS TO ASSIGNMENT 01/2008
 2

SOLUTIONS TO ASSIGNMENT 01

Dear Student,
The solutions of assignment 01 follow below. Please study the solutions carefully together with your marked
assignment.

1. Use the operator method (method of elimination) to solve the following systems of differential equa-
tions:

(a) (6D − 2)[x] + (D + 3)[y] = 0 (1)

t
(4D − 6)[x] + (3D + 2)[y] = 7e . (2)
N.B. Solve first for x. Use the method of undetermined coefficients when solving for y.
(b) 2ẋ + x + ẏ − y = −3 cos t
ẋ + 2x + ẏ + 3y = 5 sin t
N.B. Eliminate y first.

Solution:

(a) (3D + 2)[(1)] : (18D 2 + 6D − 4)[x] + (D + 3)(3D + 2)[y] = 0 (3)

2 t
(D + 3)[(2)] : (4D + 6D + 18)[x] + (D + 3)(3D + 2)[y] = 28e (4)

(3) − (4) : (D 2 + 1)[x] = −2et .

The auxiliary equation is m 2 + 1 = 0. Hence m = ±i which yields the complementary function

xC.F. (t) = A cos t + B sin t.

The particular integral is

−2et −2et
x P.I. (t) = 2
= 2
= −et ,
(D + 1) (1 + 1)
1
by using B(1), i.e. the first property of the inverse operator (see Tutorial Letter 102).
D2 + 1

Therefore

x(t) = xC.F. (t) + x P.I. (t)

= A cos t + B sin t − et

with A and B arbitrary constants.

Substitution in (1) yields

(D + 3)[y] = −(6D − 2)[x] = −(6D)[x] + 2x

= 4et + (2A − 6B) cos t + (6A + 2B) sin t.
 3 MAT306X/201

For this the auxiliary equation is m + 3 = 0, i.e. m = −3, which yields the complementary function

yC.F. (t) = Ge−3t , G an arbitrary constant.

We use the method of undetermined coefficients to find the particular integral y P.I. (t) :
Suppose
y P.I. (t) = Cet + D cos t + E sin t,

then

(D + 3)[y P.I. ] = 4Cet + (3D + E) cos t + (3E − D) sin t

= 4et + (2A − 6B) cos t + (6A + 2B) sin t.

Comparing the coefficients of et , cos t and sin t, we find

et : 4C = 4 ⇒ C = 1
cos t : 3D + E = 2A − 6B (5)
sin t : −D + 3E = 6A + 2B (6)

Solving D and E in terms of A and B, it follows from (5) and (6) that

E = 2A and D = −2B.

Hence
y P.I. (t) = et − 2B cos t + 2A sin t

so that

y(t) = yC.F. (t) + y P.I. (t)

= Ge−3t + et − 2B cos t + 2A sin t.

Since the system is of order 2, there are too many constants. Substitution into (2) shows that G = 0. The
complete solution of the system is therefore

x(t) = A cos t + B sin t − et

y(t) = 2A sin t − 2B cos t + et .

(b) In operator notation the system becomes

(2D + 1)[x] + (D − 1)[y] = −3 cos t (1)

(D + 2)[x] + (D + 3)[y] = 5 sin t (2)
 4

(D + 3)[(1)] : (2D 2 + 7D + 3)[x] + (D + 3)(D − 1)[y]

= 3 sin t − 9 cos t (3)
(D − 1)[(2)] : (D 2 + D − 2)[x] + (D + 3)(D − 1)[y]
= 5 cos t − 5 sin t (4)

(3) − (4) : (D 2 + 6D + 5)[x] = 8 sin t − 14 cos t

from which
xC.F. (t) = Ae−5t + Be−t

8 sin t 14 cos t
x P.I. (t) = − 2
D2 + 6D + 5 D + 6D + 5

8 sin t 14 cos t
= −
(−12 ) 2
+ 6D + 5 (−1 ) + 6D + 5

8 sin t 14 cos t
= −
6D + 4 6D + 4

by using B(3) of Tutorial Letter 102. Now from B(5) of Tutorial Letter 102 with λ = 6, µ = 4 and
a = 1, it follows that
8 sin t 4
=− (3 cos t − 2 sin t).
6D + 4 13
Similarly
14 cos t 7
− = − (3 sin t + 2 cos t).
6D + 4 13
Therefore

x(t) = xC.F. (t) + x P.I. (t)

4
= Ae−5t + Be−t − 13
(3 cos t − 2 sin t)
7
− 13 (3 sin t + 2 cos t)

= Ae−5t + Be−t − 2 cos t − sin t (5)

By substituting in (2) we get

(D + 3)[y] = 3Ae−5t − Be−t + 5 cos t + 5 sin t, (6)

so that clearly, from the auxiliary equation m + 3 = 0,

yC.F. (t) = Ce−3t .

Put
y P.I. (t) = De−5t + Ee−t + F cos t + G sin t.
 5 MAT306X/201

Substitution in (6) yields

−2De−5t + 2Ee−t + (3G − F) sin t + (3F + G) cos t

= 3Ae−5t − Be−t + 5 cos t + 5 sin t.

By comparing coefficients of e−5t , e−t , sin t and cos t we get

3 1
D = − A, E = − B, F = 1, G = 2.
2 2
It therefore follows that
3 1
y(t) = − Ae−5t − Be−t + cos t + 2 sin t + Ce−3t .
2 2
Substitution in (1) yields C = 0, so that
3 1
y(t) = − Ae−5t − Be−t + cos t + 2 sin t. (7)
2 2
The complete solution is now given by (5) and (7).

2. Triangularize the following system and determine the solution

(D − 2)[y1 ] + 2D[y2 ] = 2 − 4e2x (1)

(2D − 3)[y1 ] + (3D − 1)[y2 ] = 0. (2)

N.B. Reduce to the form

P1 (D)y2 = f 1 (x)
P2 (D)y1 − P3 (D)y2 = f 2 (x).

Solution:
2(1) − (2) : −y1 + (D + 1)[y2 ] = 4 − 8e2x (1 )
(2D − 3)[y1 ] + (3D − 1)[y2 ] = 0 (2)

Now substitute
y1 = (D + 1)[y2 ] − 4 + 8e2x

in (2). This yields the triangular system

(D 2 + D − 2)[y2 ] = −6 − 4e2x (3)

2x
y1 − (D + 1)[y2 ] = −4 + 8e (4)

The auxiliary equation for (3) is

m 2 + m − 2 = (m + 2)(m − 1) = 0,
 6

i.e. m = −2 or m = 1. Hence

y2C.F. (x) = Ae−2x + Be x

−6e0x −4e2x
y2 P.I. (x) = +
D2 + D − 2 D2 + D − 2

−6e0x −4e2x
= +
02 + 0 − 2 22 + 2 − 2

= 3 − e2x

1
having made use of property B(1) of the inverse operator (see Tutorial Letter 102). This gives
f (D)

y2 (x) = y2C.F. (x) + y2 P.I. (x)

= Ae−2x + Be x + 3 − e2x (5)

Hence, from (4) it follows that

y1 (x) = −4 + 8e2x + (D + 1)[y2 ]

= −Ae−2x + 2Be x + 5e2x − 1 (6)

The complete solution is now given by (5) and (6).

3. Decide whether the following systems have no solutions or infinitely many solutions. Find the general
form of the solution in each case.

(a) ẏ1 + ẏ2 = 3

ÿ1 − ẏ1 + ÿ2 − ẏ2 = x

(b) ẏ1 − y1 + ẏ2 = x

ÿ1 − ẏ1 + ÿ2 = 1

Solution:

(a) In operator notation the system is

D[y1 ] + D[y2 ] = 3
D(D − 1)[y1 ] + D(D − 1)[y2 ] = x
 7 MAT306X/201

Hence
= D 2 (D − 1) − D 2 (D − 1) = 0.

But since
3 D
2
= (D 2 − D)[3] − D[x] = −1 = 0,
x D −D
it follows that this system has no solution. (This can also be seen by applying D−1 to the first equation.)
(b) In operator notation the system is

(D − 1)[y1 ] + D[y2 ] = x
(D 2 − D)[y1 ] + D 2 [y2 ] = 1.

In this case
= D 2 (D − 1) − D(D 2 − D) = 0.

Furthermore, since
x D
= D 2 [x] − D[1] = 0
1 D2
and
D−1 x
= (D − 1)[1] − (D 2 − D)[x] = −1 + 1 = 0,
D2 − D 1
the system has infinitely many solutions. These solutions may be obtained from any of the two given
equations. Let for instance, y1 = g(x) where g(x) is an arbitrary function. Then, from the first equation,

y2 (x) = x − g (x) + g(x)

whence
x2
y2 (x) = − g(x) + g(x)d x + c
2
where c is an arbitrary constant.

4. Use the eigenvalue–eigenvector method to solve the initial value problem

   
−1 1 0 7
   
Ẋ =  1 2 1  X, X(0) =  2  .
0 3 −1 5

Solution:
The characteristic equation C(λ) = −(1 + λ)(λ + 2)(λ − 3) has roots λ = −2, λ = −1 and λ = 3.

λ = −2 : Solve     
1 1 0 v1 0
    
 1 4 1   v2  =  0  ,
0 3 1 v3 0
 8

i.e. the system

v1 + v2 = 0 (1)
v 1 + 4v 2 + v 3 = 0 (2)
3v 2 + v 3 = 0 (3)

From (1) we have v 1 = −v 2 and from (3), v 3 = −3v 2 . Hence, choosing v 2 = 1, we find v 1 = −1 and
v 3 = −3. An associated eigenvector is then
 
−1
 
V1 =  1  .
−3

In the same way we find the eigenvectors


 

1 1
   
V2 =  0  and V3 =  4 
−1 3

corresponding to λ = −1 and λ = 3 respectively. The general solution therefore is

     
−1 1 1
  −2t   −t   3t
X(t) = c1  1  e + c2  0  e + c3  4  e
−3 −1 3

with c1, c2 and c3 arbitrary constants.

From the initial condition 
7
 
X(0) =  2 
5
we find

−c1 + c2 + c3 = 7 (4)
c1 + 4c3 = 2 (5)
−3c1 − c2 + 3c3 = 5 (6)

Then

(4) + (5) : c2 + 5c3 = 9 (7)

3 × (5) + (6) : −c2 + 15c3 = 11. (8)

(7) + (8) gives c3 = 1 so that c2 = 4. Hence from (4) follows that c1 = −2.
The solution is therefore
     
−1 1 1
  −2t   −t   3t
X(t) = −2  1  e + 4  0  e +  4  e .
−3 −1 3
 9 MAT306X/201

5.(a) Show from first principles, i.e. by using the definition of linear independence, that if λ = a+ib, b = 0 is
an eigenvalue of a real matrix A with associated eigenvector v = u+iw, then the two real solutions

Y(t) = eat (u cos bt − w sin bt)

and
Z(t) = eat (u sin by + w cos bt)

are linearly independent solutions of Ẋ = AX.

(b) Solve the system

6 −1
Ẋ = X.
5 2

N.B. Real solutions are required.

Hint: Recall that if the roots of C(λ) = 0 occur in complex conjugate pairs, then any one of the
two eigenvalues yields two linearly independent solutions. The second will yield solutions which are
identical (up to a constant) to the solutions already found. Check Example 2.2 of the Study Guide in
this regard.

Solution:

(a) Using first principles, we start with the assumption

c1 Y(t) + c2 Z(t) = 0 for all t. (1)

Recalling the definition of linear independence, we have to show c1 = c2 = 0. Since (1) holds for all t, we
π
can conclude, by taking t = 0 and t = 2b in (1), that

c1 u+c2 w = 0 (2)

aπ
e 2b (−c1 w+c2 u) = 0. (3)

This gives
c1 u = −c2 w and c1 w =c2 u,

so that
c12 u = −c1 c2 w = −c2 c1 w = −c22 u (4)

and
c12 w = c1 c2 u = c2 c1 u = − c22 w. (5)

Since v = u+iw is not the zero vector, either u or w (or both) is not the zero vector. If u = 0, then from (4)
we obtain
c12 = −c22 .
 10

The left hand side of this equation is non–negative, while the right hand side is non–positive. Hence c1 =
0 and c2 = 0 are the only values which satisfy the equation. If w = 0, then a similar argument using equation
(5) shows that c1 = 0 and c2 = 0. The solutions Y and Z are therefore linearly independent.

(b) The characteristic equation has roots λ = 4 ± i. To find the eigenvector corresponding to 4 + i, we solve the
system
2−i −1 v1 0
=
5 −2 − i v2 0

which is equivalent to

(2 − i)v 1 − v2 = 0 (1)
5v 1 − (2 + i)v 2 = 0. (2)

(2 − i)
(Note that − × (2) = (1), i.e. the second equation is identical to the first one.) Letting v 1 = 1, we get
5
v 2 = 2 − i, so that an eigenvector corresponding to 4 + i is

1 1 0
V= = +i .
2−i 2 −1

Using the formulae for two real solutions in 5(a), we get

1 0
Y1 (t) = e4t cos t − sin t
2 −1

and
0 1
Y2 (t) = e4t cos t + sin t .
−1 2

The general solution is then

Y(t) = c1 Y1 (t) + c2 Y2 (t), c1 , c2 arbitrary constants.

I trust that you will find the solutions useful and are looking forward to your next assignment.

Your MAT306X lecturer.