MTH6141 Random Processes, Spring 2012, Exercise Sheet 7

Please drop your answer to Questions 2 and 4 in the red box on the second oor of the maths building by 13:00 on Tuesday 13th March 2012. You are strongly encouraged to attempt as many questions as possible. Please send comments and corrections to m.jerrum@qmul.ac.uk. 1. (a) I arrive at a bus stop at a random time. What is my expected waiting time under the following assumptions: i. I am in Z urich, and the buses are equally spaced in time, 10 minutes apart. ii. I am in London, and the buses arrive according to a Poisson process with rate 6 per hour? (b) How would you explain the dierence in your answers to the two parts to a non-mathematician? (Note that the expected number of buses per hour is the same, namely 6, in both cases.) 2. Complete the proof of Theorem 2.2. In lectures, we saw that FT1 (t) (the c.d.f. of T1 ) is FT1 (t) = P(X (t) 1) = 1 et , and by dierentiation that fT1 (t) (the p.d.f. of of T1 ) is et . We also saw, likewise, that FT2 (t) = P(X (t) 2) = 1 et tet , and fT2 (t) = 2 tet . By considering the general case, show that fTn (t) = n tn1 t e , (n 1)!

i.e., that Tn has the Gamma distribution. 3. Let X (t) be a Poisson process of rate . Show that for all n > 0 the probability density function of T1 conditioned on X (t) = n is fT1 |X (t)=n (u) = for 0 < u t. [Hint: Work out the cdf rst.] Use this to nd E(T1 | X (t) = n). n u 1 t t
n1

4. Let T1 , T2 , . . . be the arrival times of a Poisson process X (t) of rate . Find the following (a) E(T1 + T2 + T3 | X (4) = 3), (b) E(T1 + T2 + T3 + T4 | X (4) = 3), and
2 2 2 2 2 | X (2) = 5). + T5 + T4 + T3 + T2 (c) E(T1

5. Requests arrive at a server as a Poisson process of rate per minute. Every T minutes the requests are processed regardless of how many there are (even if there are none). Suppose that processing costs k (regardless of how many requests are processed). In addition each request incurs a cost of c for each minute it waits before processing. (a) Show that the expected cost per minute to run the server is k cT + . T 2 [Hint: First work it out conditioned on there being n requests waiting at time T .] (b) How should T be chosen to minimize the cost of running the server?