ABSTRACT
A polling model is a queue system which consists of N queues, which are attended by a single server by visiting the queues in a cyclic order [1-3].
Many studies in the literature assume that customers arrive at queues follow the Poisson process, because it is tractable for performance analysis. In many applications of polling system, this assumption is appropriate, such as in computercommunication systems, and information packets arrive at the termination with the Poisson process. Yet this might in some cases, not be a realistic assumption, e.g., in bursty systems like voice and video, transportation systems, and some production systems, events might occur simultaneously. We study the system with Markova Batch arrival where the number of customers arriving at queue in each slot is governed by a Markov chain.
