Exams Civ6400 Jan05

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

CIV6400

CIV6400 1
CIV6400

Q1

Given the differential equation

d2 d
( 2
+ 2 + 5)X = f (t) (1)
dt dt
a.) Give a fundamental solution system of the homogeneous problem f (t) = 0. 5 marks

b.) Show by means of the Wronskian determinant that the system of solutions of the
homogeneous problem you gave in a.) indeed is a fundamental system. 5 marks

c.) Determine the general solution X(t) for f (t) = eα t for α 2 + 2α + 5 6= 0. 5 marks

d.) Solve the initial value problem

d2 d
( 2
+ 2 + 5)X = 2 cost (2)
dt dt
with initial conditions
X(t = 0) = − 35 ,
dX 1
dt (t = 0) = 5 .

Hint: make use of 2 cost = eit + e−it , i = −1. 10
marks

e.) (i) Consider now the non-homogeneous problem (1) for α = iβ , i.e. f (t) = eiβ t
where β ∈ R and determine for this case the so called enhancement factor V which is defined
by

|X p (t)|
V =: lim (3)
t→∞ | f (t)|

where X p (t) denotes any particular solution of the non-homogeneous problem. 10


marks

(ii) Why is definition (3) unique, i.e. independent on given initial conditions? 2 marks

(iii) Determine the exciting frequencies β for which V is a maximum and determine
the maximum value of V . 5 marks

(iv) Make a schematic drawing of V (β ). 3 marks

f.) Determine the so called phase shift for f (t) = eiβ t which is defined by

φ =: lim {arg(X p (t) − arg( f (t))} (4)


t→∞

Hint: The argument of a complex number z = x+iy = ρ eiϕ is given by arg(z) = ϕ = tan−1 ( xy ).
5 marks

CIV6400 2 CONTINUE
CIV6400

Q2

Given the recursion

Xn+1 = Xn + bYn ; Yn+1 = bXn +Yn ; n = 0, 1, 2, .. (5)

with b > 0 and initial values X0 = 1, Y0 = 0.

a.) (i) Determine the limiting case A = limn→∞ ( XYnn ) for b > 0. Hint: It is convenient
to consider the given recursion in matrix form, i.e. ~xn+1 = M~xn with start vector x~0 . 10
marks

(ii) Does A = limn→∞ ( XYnn ) depend on µ for initial conditions X0 = µ , Y0 = µ ? 3 marks

b.) Consider now the modified recursion

Xn+1 = p(Xn + bYn ); Yn+1 = p(bXn +Yn ); n = 0, 1, 2, .. (6)

for b > 0.

(i) For which values of p > 0 do the sums Sx = ∑∞ ∞


n=0 Xn , Sy = ∑n=0 Yn exist? 6 marks

(ii) Determine Sx and Sy for those values of p for which these sums exist by using the
initial condition X0 = 1, Y0 = 0. 6 marks

CIV6400 3 TURN OVER


CIV6400

Q3

a.) (i) Give a homogeneous differential equation of the form P(D)X = 0, D = dtd for

which the functions X1 (t) = e−t eit , X2 (t) = e−t e−it , X3 (t) = e−t , (i = −1) constitute a
fundamental system. 10
marks

(ii) Does the set of functions X̃1 (t) = e−t cost, X̃2 (t) = e−t sint, X̃3 (t) = e−t also con-
stitute a fundamental system to this homogeneous differential equation P(D) = 0? Hint:
Consider the Wronskian determinant of this set X̃1 (t), X̃2 (t), X̃3 (t). 5 marks

b.) (i) Consider this non-homogeneous differential equation P(D)X = 2 cos ω t (ω ∈ R)


with the disturbance function 2 cos ω t and determine its general solution. 5 marks

(ii) Determine the periodic function X p (t) which dominates any general solution X(t)
for large t. 5 marks

CIV6400 4 END OF PAPER


CIV6400

Suggested solution to Q1

Remark:
d
We can write the problem (1) as P(D)X = f with differential operator D = dt and
P(D) = D2 + 2D + 5 is the characteristic polynomial in D.

a.) Fundamental system to P(D)X = 0. With ansatz Xh = eλ t we find two solutions of


the homogeneous problem X1 = eλ1t , X2 = eλ2t where λi are the zeros of the characteristic
polynomial √ √
P(λ ) = λ 2 +2λ +5√= (λ − λ1 )(λ − λ2 ) withλ1 = −1± 1 − 5 = −1−2i,λ2 = −1+ 1 − 5 =
−1 + 2i where i = −1 thus a fundamental system is
X1 (t) = eλ1t = e−t e−2it
(7)
X2 (t) = eλ2t = e−t e2it

b.) X1 = eλ1t , X2 = eλ2t constitute a fundamental system if the Wronskian determinant W 6= 0


(criteria for linear independence of the solutions of the homogeneous problem). For X1 =
eλ1t , X2 = eλ2t the Wronskian is
W (t) = (λ2 − λ1 )e(λ1 +λ2 )t = 4ie−2t 6= 0 sinceλ1 6= λ2 (8)
That is X1 = eλ1t , X2 = eλ2t is a fundamental system of P(D)X = 0 for non-multiple zeros
λ1 6= λ2 and the general solution of the homogeneous problem P(D)X = 0 then writes
(superposition principle because of linearity of the problem) Xh = C1 eλ1t +C2 eλ2t (Ci = const)

c.) The general solution of any non-homogeneous problem P(D)X = f (t) has the form

X(t) = C1 X1 (t) +C2 X2 (t) + X p (t) (9)


where X p (t) is any particular solution of the non-homogeneous problem, i.e. solving
P(D)X p = f (t)
A Particular solution of the non-homogeneous problem with f (t) = eα t for α 2 + 2α +
5 6= 0, i.e. α 6= λi can be found by putting X p (t) = Ceα t which leads to

P(D)X p (t) = CP(α )eα t = eα t (10)


αt
or C = P(1α ) and hence X p (t) = P(e α ) , with P(α ) = α 2 + 2α + 5 6= 0 is a particular solution.
Hence the general solution (9) has the form

eα t
X(t) = e−t (C1 e−2it +C2 e+2it ) + (11)
(α 2 + 2α + 5)

d.) General solution for f (t) = eit + e−it = 2Re{eit } (x p sum of particular solutions for
α = i and α = −i
eit
X(t) = e−t (C1 e−2it +C2 e2it ) + 2Re{ } (12)
(4 + 2i)
1
X(t) = e−t (C1 e−2it +C2 e2it ) + (2 cost + sint) (13)
5
The initial conditions lead to the following equations for the Ci
2 3
C1 +C2 = − − = −1 (14)
5 5

CIV6400 5
CIV6400

2 2
(−1 − 2i)C1 + (−1 + 2i)C2 = − + = 0 (15)
5 5
which are uniquely invertible because the Wronskian determinant W (t) 6= 0 is non-vanishing
and yields (λ1 = −1 − 2i, λ2 = −1 + 2i):

λ2 1 λ1 1
C1 = − = − (2 + i), C2 = = − (2 − i) (16)
(λ2 − λ1 ) 4 (λ2 − λ1 ) 4

The solution of this initial value problem is then given by (13) with (16).

e.)(i) α = iβ , β ∈ R → P(iβ ) 6= 0 and hence (11) applies to this case and writes

eiβ t
X(t) = e−t (C1 e2it +C2 e−2it ) + (17)
(−β 2 + 2iβ + 5)

Remark: As limt→∞ Xh (t) = 0 (Xh (t) decaying as e−t ) only particular solution
ei β t
X p = (−β 2 +2i β +5)
contributes to limiting cases t → ∞ of X(t).

(ii) Since any solution X(t) of the non-homogeneous problem only differs by homoge-
neous solutions Xh (t) which tends to zero as e−t , the definition of V is unique and assumes
with
X p (t) 1
lim = (18)
t→∞ f (t) (−β + 2iβ + 5)
2

the form
|X p (t)| 1
V =: lim = (19)
t→∞ | f (t)| |(−β + 2iβ + 5)|
2

which assumes with |(−β 2 + 2iβ + 5)|2 = (β 2 − 5)2 + 4β 2 = β 4 − 6β 2 + 25 = (β 2 − 3)2 + 16


thus V of (19) yields

|X p (t)| 1 1
V =: lim =p =p (20)
t→∞ | f (t)| β 4 − 6β 2 + 25 (β 2 − 3)2 + 16

√ β − 6β + 25 = (β − 3) + 16 is a minimum, i.e. for β = 3


(iii) V is a maximum when 4 2 2 2 2

corresponding to β1,2 = ± 3.

(iii) The maximum value of V then is with (20) Vmax = 14 .

CIV6400 6
CIV6400

(iv) Drawing of V (β ):

width=2.4in,height=2.4in]enhfactor.eps

Figure 1: Enhancementfactor V (β ).

X p (t) ei φ 5−β 2 −2iβ


f.) From 1
f (t) = P(iβ ) = |P(iβ )| = |P(iβ )|2
the phase shift becomes φ with (18) tan(φ ) =
Im 5−β 2 −2iβ 2β
Re ( |P(iβ )|2 ) = β 2 −5
µ ¶
−1 2β
φ = tan (21)
β −5
2

CIV6400 7
CIV6400

Suggested solution to Q2

Remark: The recursion of (6) can be written in matrix form

~xn+1 = M~xn ; n = 0, 1, 2, .. (22)

with M1 1 = 1, M12 = M21 = b, M22 = 1 (Mi j = M ji ) hence orthogonal eigenvectors vi .


(22) can be rewritten as
~xn = Mn~x0 (23)
i.e. Mn to be determined.
Remark: Since M = VDV−1 where V = (~v1 ,~v2 ) where M~v1 = λi~vi , D = diag(λ1 , λ2 ),
→ Mn = Vdiag(λ1n , λ2n )V−1 , (V−1 = 12 VT ).

a.)(i)
x~n = Vdiag(λ1n , λ2n )V−1~x0 (24)
~ = (C1 ,C2 )T = V−1~x0 thus
Put ~x0 = C1~v1 +C2~v2 hence C

x~n = C1 λ1n~v1 +C2 λ2n~v2 (25)

Xn C1 λ1n v1x +C2 λ2n v2x


= (26)
Yn C1 λ1n v1y +C2 λ2n v2y
Eigenvalues λi (characteristic equation): |M − λ E| = 0 → (λ − 1)2 − b2 = 0 → λ1 = 1 + b >
λ2 = 1 − b. Eigenvectors are: ~v1 = (1, 1)T to λ1 = 1 + b and ~v2 = (−1, 1)T to λ1 = 1 − b and
~ = 1 VT~x0 = ( 1 , − 1 )T thus (26) becomes
hence C 2 2 2

λn
2 λ1 + 2 λ2 λ1n + λ2n 1 + λ1n
1 n 1 n 2
Xn
= 1 n 1 n= n n = (27)
Yn 2 λ 1 − 2 λ 2
λ1 − λ2
λn
1 − λ2n
1

λ2n
For b > 0 we have λ1 = 1 + b > λ2 = 1 − b. Hence limn→∞ λ1n = 0 thus
Xn v1x
lim ( )= =1 (28)
n→∞ Yn v1y
(ii) this result does not depend on µ .
b.) Eigenvalues are now pλ1,2 = p(1 ± b). We obtain with (25) for ~x0 = (1, 0)T

x~n = (C1 λ1n~v1 +C2 λ2n~v2 )pn (29)


1
with C1 = 2 and C2 = − 12 .

(i) The sums Sx,y are convergent if Max|pλi | = |pλ1 | = p(1 + b) < 1 (b > 0, p > 0,
(since ∑n ρ n < ∞ |ρ | < 1 → The sums exist for p < 1+b
1
.
1
(ii) Determination of sums for p < 1+b :
Because of ∞
1
∑ (pλ1,2)n = 1 − pλ1,2 (30)
n=0
we obtain with (29) µ ¶

C1 C2
∑ ~xn = 1 − pλ1
~v1 +
1 − pλ2
~v2 (31)
n=0

CIV6400 8
CIV6400

with C1 = 12 , C2 = − 12 , λ1 = 1 + b, λ2 = 1 − b, ~v1 = (1, 1)T , ~v2 = (−1, 1)T which leads to


µ ¶
1 1 1
Sx = + (32)
2 (1 − p(1 + b)) (1 − p(1 − b))
µ ¶
1 1 1
Sy = − (33)
2 (1 − p(1 + b)) (1 − p(1 − b))

CIV6400 9
CIV6400

Suggested Solution to Q3

a.)(i) The given fundamental system indicates roots of the characteristic polynomial
P(λ ): λ1 = −1 + i, λ2 = −1 − i, λ3 = −1, hence

P(D) = (D − λ1 )(D − λ2 )(D − λ3 ) = (D + 1 − i)(D + 1 + i)(D + 1) (34)

= [(D+1)2 −i2 ](D+1) = (D+1)3 +(D+1) = D3 +3D2 +3D+1+D+1 = D3 +3D2 +4D+2


(35)
3 2 d3 d2 d
(ii) Hence P(D)X = (D + 3D + 4D + 2)X = dt 3 X + 3 dt 2 X + 4 dt X + 2X = 0 is a homoge-
neous differential equation to which {X1 (t), X2 (t), X3 (t)} is a fundamental system.

(ii) By putting ~X̃(t) = (X̃1 , X̃2 , X̃3 )T we have ~X̃(t) = A~X(t) and hence

W̃ (t) = detAW (t) (36)

where W (t) 6= 0 denotes the Wronskian of the set given in a.). Since X̃1 = 12 (X1 + X2 ) and
X̃2 = 2i1 (X1 X2 ) and X̃3 = X3 we have detA = ( 12 (−1) 1 1 i
2i − 2i 2 )1 = 8 6= 0 →
The given set of functions X̃1 (t) = e−t cost, X̃2 (t) = e−t sint, X̃3 (t) = e−t constitutes also a
fundamental system to the homogeneous differential equation P(D) = 0.
iω t
b.)(i) P(D) = eiω t + e−iω t Hence X(t) = Xh (t) + X p (t) where X p (t) = 2Re{ P(i
e
ω ) } and
P(iω ) = 2 − 3ω 2 + iω (4 − ω 2 ).
Re(P(iω )) = 2 − 3ω 2 , Im(P(iω )) = ω (4 − ω 2 ), hence
¡ ¢
eiω t (2 − 3ω 2 ) cos ω t − ω (4 − ω 2 ) sin ω t
X p (t) = 2Re{ }=2 (37)
P(iω ) (2 − 3ω 2 )2 + ω 2 (4 − ω 2 )2

(ii) Because the general solution of the homogeneous problem Xh (t) = e−t (C1 eit + C2 e−it +
C3 ) → 0 for large t >> 1 as e−t . Because the general solution of the non-homogeneous
problem X(t) = Xh (t) + X p (t) (X p (t) given by (37)), therefore (37) is the periodic function
dominating any general solution X(t) for large times t.

CIV6400 10

You might also like