Birth and Death Processes Example: A Barber Shop A small town barbershop has two barbers and an additional chair for 1 waiting customer. If a customer arrives when there are 3 customers in the shop, the customer leaves. • The customers arrive according to a homogeneous Poisson process. The average time between arrivals is 30 minutes. • The two barbers are Andy and Bob. Andy completes serving a customer at a rate of 2 per hour. Bob completes a serving a customer at a rate of 1 per hour. • Because Andy is faster than Bob, if only one customer is in the store, Andy handles the customer. Assume that inter-arrival times and service times are independent exponential random variables. (a) Draw a state diagram with possible states and corresponding birth/death rates. Solution: 2 2 2 1 0 2 2 3 3 3 E[B] = 30min = 1/2 hrs. implying rate λ = 2 per hour. Let TA = time until Andy finishes and TB = time until Bob finishes. For k = 1, D = TA ; since µA = 2 per hour, D ∼ Exp(2) For k = 2, 3, D = min{TA , TB } giving D ∼ Exp(1 + 2 = 3); since µA = 2 and µB = 1 (b) What is the (large t) probability that the shop is empty? 1 2·2 2·2·2 + + 2 2·3 3·3·2 4 8 = 1+1+ + 6 18 28 = 9 S = 1+ implying that p0 = 1 S = 9 28 (c) Let X be number of customers in the store. pk = P [X = k] = k pk = P (X = k) λ0 λ1 . . . λk−1 p0 µ1 µ2 . . . µk 0 1 9 28 9 2 28 2 1 What is the pmf of X? 9 28 2 · 2 3 9 28 3 · 4 9 (d) What is the (large t) probability that an arriving customer is turned away? An arriving customer is lost when the shop is full i.e., X = 3 p3 = P (X = 3) = 4/28 = 1/7 (e) What is the distribution of the number of customers arriving in the first 3 hours? (Name the distribution and it parameter value(s)) Let Y (t) = # of arrivals in t hours; note that the number of arrivals is not the same as the number that enter the shop. Y (t) is a homogeneous Poisson process with rate λ = 2 per hour. Thus Y (3) ∼ P oisson(λt = 2 × 3) (f) What is the probability that no customer arrives in a 3 hour time interval? What is the expected number of arrivals in 3 hours? Since Y (3) ∼ P oisson(6), P (Y (3) = 0) = 60 −6 e 0! E[Y (3)] = 6 (g) What is the distribution of the time until the 5th customer arrives? (Name the distribution and it parameter value(s); As in part (f), an arrival does not necessarily enter the barbershop) Let O5 = the time until the 5th customer arrives. O5 ∼ Erlang(5, 2) since k = 5 and λ = 2 . (h) What is the probability that the time until the 5th customer arrives is less than 2.5 hours? From above, P (O5 ≤ 2.5) = FX (2.5) where X ∼ Erlang(5, 2) FX (2.5) can be computed using the Poisson cdf, since FX (2.5) = 1 − FY (5 − 1) where Y ∼ P oi(2 × 2.5) i.e., we need to look up the Poisson table for FY (4) where Y ∼ P oi(5). From the tables, this is equal to .44 so the required probability is 1 − .44 = .56. 2