If this happens then we have at most 3 green arrivals

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ication of this consider Example 2.4. M/G/1 queue. In modeling telephone tra c, we can, as a first approximation, suppose that the number of phone lines is infinite, i.e., everyone who tries to make a call finds a free line. This certainly is not always true but analyzing a model in which we pretend this is true can help us to discover how many phone lines we need to be able to provide service 99.99% of the time. The argument for arrivals at the Great Hall implies that the beginnings of calls follow a Poisson process. As for the calls themselves, while many people on the telephone show a lack of memory, there is no reason to suppose that the duration of a call has an exponential distribution. So we use a general distribution function G with G(0) = 0 and mean µ. Suppose that the system starts empty at time 0. The probability a call started at s has ended by time t is G(t s), so using Theorem 2.12 the number of calls still in progress at time t is Poisson with mean Zt Zt (1 G(t s)) ds = (1 G(r)) dr s=0 r =0 Letting t ! 1 and usi...
View Full Document

This document was uploaded on 03/06/2014 for the course MATH 4740 at Cornell.

Ask a homework question - tutors are online