SMSL/2024

THE UNIVERSITY OF HONG KONG

DEPARTMENT OF STATISTICS AND ACTUARIAL SCIENCE

STAT2901 Probability and Statistics: Foundations of Actuarial Science

Assignment 3

1. Let X be a random variable distributed with the following mass function f (x):

x 1 2 3 4 5 6
f (x) 0.2 0.1 0.1 0.1 0.1 0.4

Calculate E[X], E[X 2 ], Var(X), E[f (X)] and Median(X), which is defined to be inf{x : P(X ≤
x) > 0.5}.

2. Let X be a positive random variable distributed with the pdf

f (x) = Cx 1 {0 ≤ x ≤ 1/2} + (3/2)x−4 1 {x ≥ 1},

for some constant C > 0, where 1 {·} denotes the indicator function.

(a) Calculate C.
(b) Calculate the mean and standard deviation of X.
(c) Calculate the median of X, defined to be inf{x : P(X ≤ x) > 0.5}.
(d) Calculate the interquartile range of X.
(e) Without deriving the pdf of Y = min{X, 1/X}, calculate E[Y ].

3. Let t ∈ (0, 1) be a fixed constant. Write down the expectations and variances of tX for the
following cases:

(a) X has a Poisson (λ) distribution.

(b) X has a Binomial (n, p) distribution.
(c) X has a geometric (p) distribution, that is,

P(X = x) = (1 − p)x p for x = 0, 1, 2, . . . .

1
 (d) X has an exp (λ) distribution.
(e) X has a U [−1, 1] distribution.

4. Let X, Y be two discrete random variables with their joint mass function f (x, y) tabulated
below:

x\y 1 2 3 4
1 0.04 0.10 0.10 0.06
2 0.10 0.10 0.05 0.05
3 0.10 0.10 0.10 0.10

(a) Tabulate the joint cdf F (x, y) of (X, Y ) for x = 1, 2, 3 and y = 1, 2, 3, 4.

(b) Calculate the marginal mass functions of X and Y respectively. Are X and Y independent?
(c) Calculate the marginal cdf’s of X and Y respectively.
(d) Calculate the covariance and correlation coefficient between X and Y .
816 −4−2
(e) Calculate the correlation coefficient between −X − 2Y + 24 and 2X + Y − 16−8 .
(f) How does the linear relationship between X and Y compare with that between −X −
816 −4−2
2Y + 24 and 2X + Y − 16−8 ?

5. Consider the following joint density function of the pair of random variables (X, Y ):

x2 − 2ρxy + y 2
 
1
fX,Y (x, y) = p exp − , x, y ∈ (−∞, ∞),
2π 1 − ρ2 2(1 − ρ2 )

where ρ ∈ (−1, 1) is a given constant.

(a) Verify that fX,Y is a joint density function.

(b) Determine the marginal density functions of X and Y respectively.
(c) Calculate the covariance between X and Y .
(d) Deduce from (c) that X and Y are independent if and only if they are uncorrelated.

6. Consider the joint density function

f (x, y) ∝ e−x−y 1 {0 < x < y < ∞}.

(a) Find the exact formula for f .

2
 (b) Let X and Y denote random variables with joint density function f . Determine their
marginal density functions.
(c) Are X and Y as defined in (b) independent? Why?
(d) Calculate the covariance between X and Y .
(e) Let Z be a random variable independent of (X, Y ), and have the density function

fZ (z) = e−z 1 {z ≥ 0}.

(i) Calculate P 2 min{X, Z} ≥ max{Y, Z} .
 
(ii) Calculate E min{2X, Y, Z} .

7. Let X1 , X2 , . . . , Xn be n discrete random variables.

(a) Suppose that X1 , X2 , . . . , Xn satisfy

n
Y
P(X1 = x1 , X2 = x2 , · · · , Xn = xn ) = P(Xj = xj )
j=1

for all real numbers x1 , x2 , . . . , xn . Show that X1 , X2 , . . . , Xn are independent.

(b) Suppose that X1 , X2 , . . . , Xn satisfy

P(Xi = x, Xj = y) = P(Xi = x) P(Xj = y)

for i 6= j and for any real numbers x and y. Are X1 , X2 , . . . , Xn independent?

8. There are two cashiers, A and B say, in a supermarket. Time taken by either cashier to serve
one customer is exponentially distributed with rate θ > 0, independent of each other. Suppose
now that A is serving one customer and B is free. There are no queues in the supermarket. Mrs.
Lee and Mrs. Chan, after finishing their shoppings, then walk towards the counters. Suppose
that Mrs. Lee arrives at the counters first and so goes to the free cashier B naturally. If you
were Mrs. Chan, which cashier would you prefer to queue for?

9. Arrivals of buses at a bus stop follow a Poisson process of arrival rate λ > 0. Each bus has a
probability p of being full when arriving at the stop, independent of other buses. A traveller at
the bus stop wishes to catch the next non-full bus, provided it arrives within τ units of time.
He decides to catch a taxi once the period of τ time units expires. Assume throughout that
nobody gets off any bus at the bus stop.

3
 (a) Show that the numbers of full and non-full buses arriving within the period of τ time
units are independent Poisson random variables. Determine also the parameters of their
respective Poisson distributions.
(b) Given that a total of m full buses arrive at the bus stop within the period of τ time units,
calculate the conditional probability that the traveller has to catch a taxi.
(c) Let m be a non-negative integer. Given that the traveller does not have to catch a taxi,
show that the conditional probability that m full buses pass him before he gets on a
non-full bus is given by
∞
pm (1 − p) X e−λτ (λτ )j
qm = .
1 − e−λτ (1−p) j=m+1 j!

(d) Given that the traveller does not have to catch a taxi, show that the number of full buses
which have passed him has a conditional mean given by
X∞  p n 1 − e−λτ (1−p) − λτ (1 − p)e−λτ (1−p) o
mqm = −λτ (1−p)
.
m=0
1 − p 1 − e
P∞
Hint: Express m=0 mqm as a double sum, write mpm−1 as (d/dp)pm , then reverse the order of
differentiation and the double summation.

10. Let X be a non-negative random variable, and Ai be the event that i − 1 ≤ X < i.
∞
X ∞
X
(a) Show that (i − 1)Ii ≤ X < iIi , where Ii = 1 if Ai occurs and = 0 otherwise.
i=1 i=1
Deduce directly that
∞
X ∞
X
P(X ≥ i) ≤ E[X] < 1 + P(X ≥ i).
i=1 i=1

(b) By considering an exponential distribution for X of rate λ, deduce from (a) that for λ > 0,

λ + e−λ > 1 and (λ + 1)e−λ ≤ 1.

11. (a) Prove the Markov’s inequality, namely

P(|X| ≥ ε) ≤ ε−1 E |X| ,

 

for any random variable X and any constant ε > 0.

[Hint: consider the inequality |X| ≥ ε1{|X| ≥ ε}, which is always true, where 1{·} denotes the
indicator function.]

4
 (b) Let Y1 , Y2 , . . . be independent and identically distributed random variables such that
E[Y1 ] = 0 and E[Y1r ] is finite for r = 2, 3, 4. Define Ȳn = n−1 ni=1 Yi , for n = 1, 2, . . . .
P

(i) Using Markov’s inequality, show that, for any n = 1, 2, . . . ,

P |Ȳn | > n−1/6 ≤ n−10/3 n E[Y14 ] + 3n(n − 1)(E[Y12 ])2 .



[Hint:
X X X
(Y1 + · · · + Yn )4 = Yi4 + 4 Yi Yj3 + 3 Yi2 Yj2
i i6=j i6=j
X X
+6 Yi Yj Yk2 + Yi Yj Yk Y` .]
i,j,k distinct i,j,k,` distinct

P∞
(ii) Using (i) and the fact that n=1 n−α < ∞ for any constant α > 1, show that
∞
[ 
|Ȳn | > n−1/6

lim P = 0.
k→∞
n=k

Hence show that ∞ \

∞
[ 
|Ȳn | ≤ n−1/6

P = 1.
k=1 n=k


[Note: The last result implies the famous Strong Law of Large Numbers: P lim Ȳn = 0 = 1.]
n→∞

12. One gloomy day in 1982 a boat carrying N asylum-seeking married couples left a port near
Haiphong (a place in northern Vietnam) and headed for Hongkong.

(a) It was known that the Hongkong government was prepared to take in a total of M (< 2N )
refugees and repatriate all those who were in excess of the quota. Assume that the M
refugees who were granted asylum were selected randomly.
(i) What is the expected number of repatriated couples?
(ii) What is the expected number of couples who became separated as a result of this
selection procedure?
(b) Assume now that the Hongkong government adopted a new policy under which each
Vietnamese arrival had a right to try his/her fortune by tossing a fair coin once. If the
coin came up “heads”, the refugee would be granted asylum; otherwise he/she would be
repatriated. What are your answers to questions (i) and (ii) in part (a)?
(c) Find the ranges of M for which the new policy was “more humanitarian”,

5
 (i) in the sense that more refugees were expected to stay in Hongkong under the new
policy;
(ii) in the sense that fewer couples were expected to separate under the new policy.

13. Princess Aurora (better known as the Sleeping Beauty) falls asleep under the spell of Fairy
Carabosse. All it takes to awaken Aurora is a magical kiss of a prince. A group of N princes
are prepared to offer their kisses. Among them is Prince Désiré, the only person whose kiss
can awaken Aurora.
Princes are to be selected at random to kiss Aurora successively until she is awakened. The
court decides to reward each selected prince with $1 for his kiss, whether or not it is successful,
and considers the following two options for the selection procedure:

Option 1. Any unsuccessful princes are kept entertained in a luxurious palace, on which the
court has to spend $1 per prince for entertainment, thus preventing the unsuccessful
princes from queuing up again to kiss Aurora and claim the $1 rewards repeatedly.
Option 2. Any unsuccessful princes are allowed to return to the reception chamber and wait
for selection again. As a result no entertainment expenses will be incurred but then each
time the N princes are equally probable to be selected.

By considering the expected expenses incurred by the two options respectively, advise the court
on an economical choice.

14. Consider the following table which records the numbers of male and female customers in a
department store over a two-week period.

Day 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Male 7 0 3 6 0 4 23 1 10 18 6 1 4 4
Female 22 3 6 13 6 11 99 9 32 58 12 5 23 12

Let X and Y be random variables representing respectively the numbers of male and female
customers on a single day. Denote by (xi , yi ) the observations for (X, Y ) on days i = 1, 2, . . . , 14
respectively, (that is, (x1 , y1 ) = (7, 22) etc. for the above record.)
The daily observations can be assumed to be independent, so that the probability of observing
the above figures is

P(X = x1 , Y = y1 ) P(X = x2 , Y = y2 ) · · · P(X = x14 , Y = y14 ).

6
 Study the correlation between numbers of male and female customers based on the given
observations according to the following steps.

(a) Plot the points (xi , yi ) on a two-dimensional xy-plane. Is there any noticeable relationship
between the xi and yi values?
(b) A probability model is suggested for the observed data such that the joint mass function
of (X, Y ) is given by
1
P(X = xj , Y = yj ) = , j = 1, 2, . . . , 14.
14
(i) Give an expression in terms of the (xi , yi )’s for E[(X − Y )2 ]. Calculate its numerical
value.
(ii) Give an expression in terms of the (xi , yi )’s for the correlation coefficient between X
and Y under this probability model.
Calculate its numerical value. What does it tell about the relationship between X
and Y ?

15. Let V be a random variable with cdf and pdf given by F and f = F 0 , respectively. Let W be
the outcome of a Bernoulli trial such that P(W = 0) = 1 − P(W = 1) = 1/2. Assume that V
and W are independent.
Define two random variables, X and Y , as follows:

ˆ X = (V + W )/2;
ˆ if W = 0, let Y = V ; otherwise let Y = 1.

(a) Show that for any x, y ∈ R,

n
o
P(X ≤ x, Y ≤ y) = (1/2) F (2x) ∧ y + 1 {y ≥ 1} F (2x − 1) ,

where a ∧ b denotes min{a, b} and 1 {·} denotes the indicator function.

(b) Deduce from (a) the marginal cdf’s of X and Y , respectively.
(i) Is X a continuous random variable? If so, derive its pdf.
(ii) Is Y a continuous random variable? If so, derive its pdf.
(c) Express the cdf of |Y /X| in terms of F .
y
(d) Suppose that F (y) = 1 {y ≥ 0}.
y+1

7
 (i) Show that |Y /X| is a mixed random variable which is equally likely to be a constant
K and a continuous random variable Z. Specify K and the distribution of Z.


(ii) Calculate Var |Y /X| .