Search

We study a state-dependent bulk matching queueing system with impatient customers, where customers and servers visit the system from both sides. Servers provide services in a batch with a maximum size and take matching customers away instantly. To characterize such a queueing system, the corresponding Markov process is constructed by the number of complete batches of customers and the number of remaining customers in the incomplete batch. By analyzing this system, we find it difficult to obtain the joint stationary distribution of the Markov process. Therefore, we consider the tail asymptotics for the joint probabilities. Using the matrix analytic method and censoring technique, we obtain the one-term and general expansions for the nonzero elements of the rate matrices, where the coefficients of the expansions are presented in the closed form. On the basis of these expansion formulae, the exact tail asymptotic result for the joint stationary probabilities is derived. [ABSTRACT FROM AUTHOR]