Search

The queueing systems with a fixed batch size are commonly discussed in literatures. However, most real-life systems involve multiple types of servers with varying service capacities. To characterize this feature, random service capacity is considered in a double-ended queueing system in this paper, in which an arrival server can serve at most k customers with probability α k. Customers' threshold strategies are discussed based on the conditional expected waiting time and customers' beliefs in the number of servers in the orbit. Then, stationary analyses are conducted for this double-ended queueing system with random service capacity and balking customers, which includes establishing a sufficient stability condition, calculating the stationary distribution for system states and stationary performance measures. Finally, numerical examples are presented to explore how performance measures are influenced by system parameters. [ABSTRACT FROM AUTHOR]