Analysis and Comparison of Queue with N-Policy and Unreliable Server

This work considers a queueing system with N-policy and unreliable server, where potential customers arrive at the system according to Poisson process. If there is no customer waiting in the system, instead of shutting down, the server turns into dormant state and does not afford service until the number of customers is accumulated to a certain threshold. And in the working state, the server is apt to breakdown and affords service again only after it is repaired. According to whether the server state is observable or not, the numerical optimal arrival rates are computed to maximize the social welfare and throughput of the system. The results illustrate their tendency in two cases so that the manager has a strong ability to decide which is more crucial in making management decision.