Remajor Marking Scheme

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

2a) y+ Tm lss

IF Shndn e sec =-Se c

maghb hod f = se =-V)


So t sol" of he Lin m dDt is

y -sec) =SES) Ces o +C

-CosS in -

C Les
Dsing inits ond = (-.)

he
1he yo i
y = Ces Simr+
V
2 J wYkm , 9=Sinx 2Cos -Sih z

a l hmsan non y"gm )


a ) Given ) n a s o l u t i on the folloo homegeneo

inea ODE

(14) - 2x' +u=O

find the general oluion, us ing the metod reduction


Order.

>
given y)=r
one

nd op Lution
t oe
.es

&" *u"+ 2

A ( 1 + ) (*u'+2o')
-
2n
( «v'+0) +1x =O

=E

hen x w +* ° w ' +2

ot) +2 =o
markR
duted order
w' t1t) for re
equaH on

b oth odes

Snteg rati
LA

Y
Usia the method o van attion o paremeter

tind a paticuor solution the tollousi on

omo qeneou linear O DE:

C1+ " 2y' +2 =r3+.


-

part a) we t t e t u r domento

Oolution th homogune ous euahion

amd

-0 morks

for Nronskian

alow, Clx) "- 2+ 28

2 -8'+ +2
+o2

FX
+Pw) + (M

Then Pati u ar D uhon


ushe re

Fr)
CUmd
11
11

4
Question 4 :

solve the
following system of differential
equations ( without Laplace transform)
using
22 + Xz
x ! =
5m , +

'
= -8N , -
322 -

223
Zz

+ 3kg
25 = Gse , -1 4ns

Solution :

Let
( Tenby)
✗ =
.

be
system
can
Then the above

rewritten as
,

'

✗ = AX ,
where

( go i
2 1
5
A =

-3 -2

To this have to
system
solve we

find and the


the
eigenvalues
A
corresponding eigenvectors of .
The characteristic equation is

given by .

det( A -

✗I )=0

ie ,

5- ✗ 2 I

8 3- X 2 =
0
-
-
-

6 4 3- ×

get
Solving
we ,
,

Zti
✗ = 1 ,
2 - i
, .

+2

For
solving
• 7=1
,

A # u
,

get
we

F-
f- ) [:
'

seas was the

free variable
-
+ I
• For ✗ = 2- i.
solving
Au; @ its
-

we
get

÷)
fÉ+É
¥ -
+1

For Zti
• ✗ =
,
solving
Aig = + i )g

we
get


↳ =

+ I
,

the solution
general
Then is

it
(f) ,eᵗu Czech -1% get
_

✗ = C + +
, , V3 .

+ I
Marking scheme of Question 5: [4+3 marks]

Let f (n) denote the n-th derivative of f and L denote the Lapace transform.
(1) Suppose f is a function such that for all integers n 0, f (n) is continuous function on R.
Moreover, assume that there exists M > 0, K > 0 satisfying |f (n) (t)|  M eKt . Using induction
prove that for n 1,
L(f (n) )(s) = sn L(f )(s) sn 1 f (0) sn 2 f (1) (0) ··· f (n 1)
(0).
(2) Compute the Laplace transform L(f ) where f is given by f (t) = t sin(2t).

Solution of Part (a):


[1 mark] For n = 1 the statement is L(f 0 )(s) = sL(f )(s) f (0), which we prove now. By definition of
Laplace transform and doing integration by parts:
R1
L(f 0 )(s) = 0 e st f 0 (t)dtR
1
= e st f (t)|1
0 + 0 se
st
f (t)dt
[1 mark] Since |f (t)|  M eKt is given, we get limt!1 e st
f (t) = 0 for s > K. Then,
0
L(f )(s) = sL(f )(s) f (0).
We have verified the case n = 1.
[1 mark] Now assume that the statement is true for n = k, i.e.,
L(f (k) )(s) = sk L(f )(s) sk 1 f (0) sk 2 f (1) (0) ··· f (k 1)
(0).
Now we claim that the statement is true for n = k + 1.
[1 mark] Consider
L(f (k+1) )(s) = L((f (k) )0 )(s)
= s(L(f (k) )(s) f (k) (0) (from the case n = 1 since |f (k) (t)|  M eKt )
= s sk L(f )(s) sk 1 f (0) sk 2 f (1) (0) · · · f (k 1) (0) f (k) (0)
= sk+1 L(f )(s) sk f (0) sk 1 f (1) (0) · · · f (k) (0).
Thus, the statement for n = k + 1.
Therefore by principle of mathematical induction, the given statement is true for all n 1.

Solution of Part (b):


d d
[1 mark] Recall that L(tf (t))(s) = ds
L(f (t)(s). Using this we get L(t sin 2t)(s) = ds
L(sin 2t)(s).

[2 mark] Using definition and simple computation gives


Z 1
2
L(sin 2t)(s) = e st t sin(2t)dt = · · · = .
0 s2 +4
Therefore, ✓ ◆
d 2 4s
L(tf (t))(s) = 2
= .
ds s +4 (s2 + 4)2
MTL 101, RE-MAJOR, PROBLEM 6

Solve the following IVP using the Laplace Transform:


y 00 + 2y = g(t) + (t ⇡/2); y(0) = 1, y 0 (0) = 0

(
0 if t < ⇡/2
g(t) =
sin t if t > ⇡/2
Solution. Apply Laplace transform to get
L(y 00 ) + 2L(y) = L(g(t) + L( (t ⇡/2))
s2 L(y) sy 0 (0) y 0 (0) + 2L(y) = L(g(t) + L( (t ⇡/2))
(s2 + 2)L(y) = s + L(g(t)) + L( (t ⇡/2))
s L(g(t)) L( (t ⇡/2))
L(y) = + 2 +
s2+2 s +2 s2 + 2
✓ ◆
1 s L(g(t)) L( (t ⇡/2))
y(t) = L + 2 +
+2 s2 s +2 s2 + 2
✓ ◆ ✓ ◆ ✓ ◆
1 s 1 L(g(t)) 1 L( (t ⇡/2))
=L +L +L
s2 + 2 s2 + 2 s2 + 2
✓ ◆ ✓ ◆
p 1 L(g(t)) 1 L( (t ⇡/2))
= cos 2t + L +L
s2 + 2 s2 + 2
✓ ⇡s/2 ◆ ✓ ◆
p 1 e 1 L(g(t))
= cos 2t + L +L
s2 + 2 s2 + 2
✓ ◆ ✓ ◆
p 1 ⇡s/2 1
p 1 1 p
= cos 2t + L e p L(sin 2t) + L L(g(t) p L(sin( 2t)
2 2
p 1 ⇣ p ⌘ 1 ⇣ p ⌘
= cos 2t + p L 1 L(u(t ⇡/2) sin 2(t ⇡/2)) + p L 1 L(g(t))L(sin( 2t)
2 2
p 1 p 1 p
= cos 2t + p u(t ⇡/2) sin 2(t ⇡/2) + p g(t) ⇤ sin( 2t)
2 2
Now computing the convolution we get
p Z t p
g(t) ⇤ sin( 2t) = g(⌧ ) sin( 2(t ⌧ ))d⌧
(0
0 t < ⇡/2
= p p p
2 sin t 2 cos( 2(t ⇡/2)) t > ⇡/2

1
Marking scheme of Question 8: [3+3 marks]

Prove or disprove the following statements.


(a) There exists A 2 M3⇥3 (R) such that the subset {I, A, A2 , A3 } is linearly independent in the vector
space M3⇥3 (R) over R.
(b) Let V be a finite dimensional vector space over R. If T, S : V ! V are linear transformations
satisfying
ker(T ) + ker(S) = V = Range(T ) + Range(S),
then
ker(T ) \ ker(S) = {0} = Range(T ) \ Range(S).
[Notation: ker(T ) and Range(T ) denote the kernel (i.e., the null space) of T and the range of T .]

Solution of Part (a) : The statement is NOT TRUE.

[1 mark] By Cayley-Hamilton theorem every matrix satisfies its characteristic polynomial.

[1 mark] For any 3 ⇥ 3 matrix A, its characteristic polynomial has degree equal to 3. Therefore, A will
satisfy a degree 3 polynomial, say x3 + a2 x2 + a1 x + a0 .

[1 mark] Then A3 + a2 A2 + a1 A + a0 I = 0, which is a non-trivial linear combination of I, A, A2 , A3


(because the coefficient of A3 is 1) which equals 0. Therefore the given subset is linearly dependent for
any A.

Solution of Part (b) : The statement is TRUE.


Write n(T ) and r(T ) for the nullity of T and the rank of T respectively. Similarly for S.

[1 mark] Given ker(T ) + ker(S) = V = Range(T ) + Range(S) implies


n(T ) + n(S) dim(ket(T ) \ ker(S)) = dim V = r(T ) + r(S) dim(Range(T ) \ Range(S)).
[1 mark] On the other hand, by rank-nullity theorem we have
r(T ) + n(T ) = dim V = r(S) + n(S).
This give us
r(T ) = n(S) dim(ket(T ) \ ker(S)) and n(S) = r(T ) dim(Range(T ) \ Range(S))
) dim(ket(T ) \ ker(S)) + dim(Range(T ) \ Range(S)) = 0.
[1 mark] Since dimension is always non-negative integer, we get
dim(ket(T ) \ ker(S)) = 0 = dim(Range(T ) \ Range(S)) = 0.
This implies
ket(T ) \ ker(S) = {0} = Range(T ) \ Range(S).

You might also like