An \(\text{M}^{\text{X/GI}}/1/\text{N}\) queue with close-down and vacation times (Q1305823)

This paper deals with an \(\text{M}^{\text{X}}/\text{GI/1}/N\) type queue, where \(N\) is the number of waiting spaces, including the one in service. Customers accepted by the system are served by a single server exhaustively (i.e., continuously until the queue is empty) under FIFO service discipline. Whenever the queue becomes empty, a close-down period, with arbitrary distribution, starts. During this period, if a customer arrives, the server immediately begins service. If not customer arrives till the end of close-down period, the server starts a vacation and if the queue is empty upon his return, he takes another vacation. Otherwise, he starts service. The vacation periods are independently and identically distributed with an arbitrary distribution function. This system is studied under two types of batch acceptance strategies in case an arriving batch size is larger than the available free waiting space. In the first, the excess number of customers will be lost (partial batch acceptance strategy) and in the second the arriving batch itself will be lost (the whole batch acceptance strategy) and here it is assumed that there is a positive probability that the batch size does not exceed \(N\). Introducing the remaining service time (close-down time, vacation time, respectively) as supplementary variable, the authors obtain a set of differential equations and show how to solve these equations to obtain the queue length at an arbitrary time as well as at a departure epoch. Further, the loss probabilities and Laplace transform of the waiting time distributions for the first customer and for an arbitrary customer are obtained. Numerical examples, to illustrate the influence of the various distributions on loss probability and the mean waiting time, are also provided.