NATIONAL UNIVERSITY OF SCIENCE AND TECHNOLOGY

SORS4103

FACULTY OF APPLIED SCIENCES

DEPARTMENT OF STATISTICS AND OPERATIONS RESEARCH

SORS4103: STOCHASTIC PROCESSES

OCTOBER 2019 TUTORIAL WORKSHEET I

Time : N/A hours

Candidates should attempt any ALL questions. All questions carry equal marks.

A1. What do you understand by the following terms

(a) Stochastic process [2]

(b) Persistent null state [2]
(c) Ergodic Markovian chain [2]

A2. Customers arrive at a bank at a Poisson rate λ. Suppose two customers arrived during
the first hour. What is the probability that

(a) both arrived during the first 20 minutes? [4]

(b) at least one arrived during the first 20 minutes? [5]

A3. A house wife buys three kinds of cereals A, B and C. She never buys the same cereal
on successive week. If she buys A, the next week she buys B or C, then the next week,
she is likely to buy A as the other brand. Find the transition matrix. In the long run,
how often does she buy each of the three brands? [5]

page 1 of 5
 SORS4103

A4. An elevator in an apartment building serves floors B, G, 1st , 2nd , 3rd , 4th , 5th , 6th , 7th , 8th
and Roof, but is malfunctioning. It stops at every floor and is twice as likely to go down
as up. A cat Burglar enters the elevator on the 4th floor.She has a waiting helicopter
on the roof and the police are waiting for her on the 2nd . What is the probability of
her escape? [5]

1
A5. For a branching process with probability Pj = 2j+1 , (j = 0, 1, 2, ...) that any individual
has j descendants in the next generation, find the mean and variance of the population
size of the nth generation. [7]

A6. On January 1 (this year) Bakery A had 40% of its local market share while the other
two bakeries B and C had 40% and 20% respectively of the market share. Based
upon a study by a marketing research firm,the following facts are compiled.Bakery
A retains 90% of its customers while gaining 5% of B’s customers and 10% of C’s
customers.Bakery B retains 85% of its customers, while gaining 5% of A’s customers
and 7% of C’s customers. Bakery C retains 83% of its customers and gains 5% of A’s
customers and 10% of B’s customers. What will be each bakery’s share on January 1,
next year, and what will be each bakery’s market share at equilibrium. [8]

A7. (a) A Markov chain {Xn , n ≥ 0} has state space E = {1, 2, 3} and transition matrix
 
0.5 0 0.5
P =  0.2 0.4 0.4  .
0.4 0 0.6

and initial distribution vector p~(0) = (0, 1, 0). Find

(i) P r{X3 = 3|X0 = 1, X1 = 1}.
(ii) P r{X3 = 3, X1 = 3}.
(iii) P r{X3 ∈ {2, 3}}.
[12]
(b) Consider a Markov chain with states E = {1, 2, 3, 4} and x
 
0.5 0.5 0 0 0
 0.5 0.5 0 0 0 
 
P = 0
 0 0.5 0.5 0  .
 0 0 0.5 0.5 0 
0.25 0.25 0 0 0.5

Determine the persistent states [8]

page 2 of 5
 SORS4103

A8. A gambler is playing a game consisting of a series of hands with her opponent. She
starts with $1 and her opponent with $31 and they play until one of them is ruined.
2
(a) On each hand the gambler wins with probability 3
and loses with probability 13 .
Which is the gambler’s better strategy,
(i) to bet $1 on each hand? [5]
(ii) to bet all she has on each hand? [5]
1 2
(b) What if the gambler’s probabilities are altered to p = 3
and q = 3
[10]

A9. A fire and emergency rescue service receives calls for assistance at a rate of per day.
Teams man the service in 12 hour shifts. Assume that requests for help form a Poisson
process

(a) from the beginning of a shift how long would the team expect to wait until their
call? [3]
(b) what is the probability that a team would receive 6 requests for help in a shift? [6]
(c) what is the probability that a team has no requests for assistance in a shift? [5]
(d) of calls for assistance, one in five is a false alarm. What is the probability that a
team has 6 requests for help in a shift but no false alarms? [6]

A10. A manufacturing company has a certain piece of equipment that is inspected at the
end of each day and classified overhauled, good, fair and imperative. If the item is
imperative it is overhauled, the procedure that takes one day. Let us denote the four
classifications as states 1,2,3 and 4 respectively. Assume that the working condition of
the equipment follows a Markov process with the following transition matrix,

0 43 41 0
 
 0 1 1 0 
P = 2 2
 0 0 1 1 .

2 2
1 0 0 0

If it costs $125 to overhaul a machine on the average and $75 in a production is lost
if a machine is found imperative. Using the steady-state probabilities to compute the
expected per day cost of maintenance. [20]

A11. Suppose disk crashes in a particular cloud computing center occur according to a
Poisson process with mean rate λ per hour, 24 hours per day, seven days a week (i.e.
24/7). Express each of the following three quantities in terms of λ, and explain or
show your work

(a) The expected number of disk crashes in a 24 hour period. [2]

(b) The probability there are exactly three disk crashes in a five hour period. [3]

page 3 of 5
 SORS4103

(c) The mean number of hours from the beginning of a particular day, until three
disks have crashed. [2]
(d) The pdf of the time the third disk crashes. [4]

A12. You are now given the result that, under the condition that n events occur in time t,
the times S1 , S2 , ..., Sn at which these n events occur are jointly distributed as indepen-
dent uniform random variables on [0, t]. In a factory manufacturing tablet computers,
assembled items arrive at the packaging point as a Poisson process of rate λ per unit
time. At a fixed time τ after the previous dispatch time, all packaged items are dis-
patched from the factory.

(a) Explain why the expected total waiting time of all the items dispatched at time
1
τ is λτ 2 . [3]
2
(b) Investment funds are available to introduce two additional dispatch times at u
and v respectively, where 0 < u < v < τ .
(i) Show that the expected total waiting time of all items dispatched at times u,
v and τ will now be
λ 2
u + (v − u)2 + (τ − v)2 .

h(u, v) = [2]
2
(ii) Deduce the optimal choice of u and v to minimise the expected total waiting
time and show that it does not depend on τ . [5]

A13. In the Ultrastratos civilisation on planet Anachronista, the population is divided into
four strata which, in order of status, are labeled Alpha, Beta, Gamma and Delta. By
the traditions of the civilisation, no child can have a status more than one different
from its parents. As examples, in each generation 20% of the children of Alphas grow
up to be Betas, the rest remaining Alphas, while of the Beta offspring 50% remain
Betas while 10% become Alphas and the rest Gammas.
A Markov chain describes the status of members of the population at successive gener-
ations. Its transition matrix is given by P, defined below, in which some entries labeled
∗ have been omitted.
Alpha  12 14 0 41 
Beta  0 12 41 14 
P = Gamma  0 0 12 12 
Delta 0 0 0 1

(a) Use the information above to fill in the missing elements of P and calculate the
two-step-ahead transition matrix. [4]
(b) What is the probability that the child of a Beta becomes a Gamma? Show that
the grandchild of a Gamma is twice as likely to become a Delta as the grandchild
of a Delta is to become a Gamma. [3]

page 4 of 5
 SORS4103

(c) Calculate the probability that, after four generations, a descendant of an Alpha
is a Delta. [3]
(d) Explain how you can tell that the chain consists of a single irreducible class. Are
any of the states transient? Give your reason. [3]
(e) Find the stationary distribution of the chain. If the population initially has no
Deltas, what will be the proportion of Deltas after a large number of genera-
tions? [7]

END OF QUESTION PAPER

page 5 of 5