Assignment 4 Solution Part A

1. Solution for Q1:

For 0 < 1, we have
Z Z 1
P (U ) = P (U |T = t) fT (t) dt = P (U |T = t) 1 dt
0
Z Z 1

= 1 1 dt + 1 dt = + ( ln t)|t=1
t= = (1 ln ) (1)
0 t
therefore, the complete CDF of U is

0 <0


FU () = (1 ln ) 0 < 1



1 1

2. Solution for Q2:

(a) For x < 0, fX (x) = 0. For x 0, we have
Z Z
fX (x) = fX,Y (x, y) dy = 2 ey dy = e x
x

We see that X is an exponential RV with E[X] = 1/. Given X = x, the conditional pdf
of Y is
fX,Y (x, y) e(yx) y > x
fY |X (y|x) = =
fX (x) 0 o.w.
To interpret this result, let U = Y X denote the interarrival time, the time between the
arrival of the first and second calls. Given X = x, then U has the same pdf as X.
(b) for y 0, Z Z y
fY (y) = fX,Y (x, y) dx = 2 ey dx = 2 y e y
0
and
fX,Y (x, y) y1 0 x < y
fX|Y (x|y) = =
fY (y) 0 o.w.
which implies that given Y = y, X is uniformly distributed in [0, y].

3. Solution for Q3:

(a) The joint characteristic function of X and Y is

X,Y (1 , 2 ) = E[ej1 X+j2 Y ] = E[ej1 X]] E[ej2 Y ]] = X (1 ) Y (2 )

j1 1) 2 22 /2
= e(e ej 2

(b) The characteristic function of Z is

Z () = E[ejZ ] = E[ej(X+Y )]] = X,Y (, )

j 1)+j 2 2 /2
= e(e

1
 Note: since Z = X + Y , and X and Y are independent, we have

Z () = X () Y ()

part (a) dealing with two-dimensional CF, while part (b) dealing with one-dimensional CF.

4. Solution for Q4:

(a) From the given joint mass function, we can have marginal pmf as
1 1 1 5
P (Y = 1) = + + =
9 3 9 9
1 1 1
P (Y = 2) = +0+ =
9 18 6
1 1 5
P (Y = 3) = 0 + + =
6 9 18
therefore, the conditional pmfs are:
P (1, 1) 1/9 1
PX|Y (X = 1|Y = 1) = = =
PY (y = 1) 5/9 5
P (2, 1) 1/3 3
PX|Y (X = 2|Y = 1) = = =
PY (y = 1) 5/9 5
P (3, 1) 1/9 1
PX|Y (X = 3|Y = 1) = = =
PY (y = 1) 5/9 5
and
1 3 1
E[X|Y = 1] = 1 +2 +3 =2
5 5 5
Similarly,
P (1, 2) 1/9 2
PX|Y (X = 1|Y = 2) = = =
PY (y = 2) 1/6 3
P (3, 2) 1/18 1
PX|Y (X = 2|Y = 2) = 0 PX|Y (X = 3|Y = 2) = = =
PY (y = 2) 1/6 3
and
2 1 5
E[X|Y = 2] = 1 +20+3 =
3 3 3
For Y = 3,
P (1, 3) 1/6 3
PX|Y (X = 1|Y = 3) = =0 PX|Y (X = 2|Y = 3) = =
PY (y = 3) 5/18 5
2
PX|Y (X = 3|Y = 3) =
5
and
3 2 12
E[X|Y = 3] = 1 0 + 2 +3 =
5 5 5
(b) From the fact that E[X|Y = 1] 6= E[X|Y = 2] 6= E[X|Y = 3], or from the fact that
PX|Y (X = 1|Y = 1) 6= PX|Y (X = 1|Y = 2) 6= PX (X = 1), we conclude that X and Y are
not independent.

2
5. Solution for Q5:
The marginal density function of Y is
Z Z y
1 y
fY (y) = f (x, y) dx = e dx = ey
0 y
therefore, the conditional density is
1 y
f (x, y) e 1
fX|Y (x|y) = = y y = (0 < x < y)
fY (y) e y
and
2 1 Zy 2 y2
E[X |Y = y] = x dx =
y 0 3
6. Solution for Q6:
(a) According to the definition of RVs, X, N and Ti , we can have
N
X
X= Ti
i=1

(b) Clearly, N is a RV with Geometric distribution with a pmf given as

k1
2 1
P [N = k] = = p q n1
3 3
1
hence, E[N ] = p
= 3.
(c) Since TN is the travel time correspondingly to the choice leading to freedom, it follows
that TN = 2 and so
E[TN ] = 2

(d) Given that N = n, the travel times Ti , i = 1, 2, , n 1 are each equally likely to be
either 3 or 5 (since we know that a door leading to the mine is selected), whereas Tn is equal
to 2 (since that choice leads to safety). Hence,
"N # "n1 #
X X
E Ti |N = n = E Ti |N = n + E [Tn |N = n]
i=1 i=1
1
= (3 + 5) (n 1) + 2 = 4n 2
2

(e)
"N # " "N ##
X X
E[X] = E Ti = E E Ti |N = E[4N 2]
i=1 i=1
= 4 E[N ] 2 = 10

3
 Assignment 4 Solution : Part B
7. Solution for Q7:
1.
(a) the state transition diagram is ignored. There are three classes: (i)state 0 is recurrent
state; (ii) state 1 and 2 are recurrent states; (iii) state 3 and 4 are transient states.
(b) From P = , we obtain the equation set:
0 + 14 3 = 0
3
+ 12 2
4 1
= 2
1
+ 12 4
4 3
= 3
1
+ 12 4
4 3
= 4
Solving the equation set, we obtain:
3
3 = 0 2 = 1 4 = 0
2
when the state transits to state 0, stationary distribution 1 = 2 = 0; when the state
transits to state 1 and 2, then from 2 = 32 1 and 1 + 2 = 1, we obtain 1 = 25 and 2 = 53 .
(c) the transition probability matrix for transient state is

1/4 1/4
PT =
1/2 1/2
therefore, 1
3/4 1/4 2.0 1.0
S = (I PT )1 = =
1/2 1/2 2.0 3.0

(d)
X
P = P (X5 = 2|X3 = 1) = P (X5 = 2|X4 = k) P (X4 = k|X3 = 1)
k
= P (X5 = 2|X4 = 1) P (X4 = 1|X3 = 1) + P (X5 = 2|X4 = 2) P (X4 = 2|X3 = 1)
3 1 1 3 9
= P12 P11 + P22 P12 = + =
4 4 2 4 16
8. Solution for Q8:
2.
Let the state be 0 of the last two trials were both successes (SS) and be 1 if last trial a
success and the one before is a failure (FS), be 2 if the last trial was a failure. The transition
matrix of this Markov chain is

0.8 0 0.2


P = 0.5 0 0.5
0 0.5 0.5

1
 Then we have equation set
0.80 + 0.51 = 0
0.52 = 1
0 + 1 + 2 = 1
we can solve that
0 + 0.40 + 0.80 = 1
which implies
5 2 4
0 = 1 = 2 =
11 11 11
Consequently, the proportion of trials that are successes is
7
0.80 + 0.5(1 0 ) =
11

10.
3. Solution for Q10:
The transition probability matrix is

0 0 0 0 1

1 0 0 0 0


P =
1/2 1/2 0 0 0


1/3 1/3 1/3 0 0

1/4 1/4 1/4 1/4 0

The equation set is (P = )

1/22 + 1/33 + 1/44 = 1

1/33 + 1/44 = 2
1/44 = 3
0 = 4
0 + 1 + 2 + 3 + 4 = 1

solving the equation, we obtain

12 6 4 3 12
= , , , ,
37 37 37 37 37

4.9. Solution for Q9:

2
3
Solutions for the questions from the textbook:

4
5
6
7
8
9
10
11