Abstract

This paper studies customers' equilibrium joining behavior in a multi-server queue with threshold policy, and then puts forward effective schemes to control customers' behaviors and to optimize social welfare. The stream of customers consists of two parts: customers in class 1 and customers in class 2, and they wait for service in order until the system is empty. Then the servers take a vacation and they would be reactivated as soon as the queue length reaches a threshold N. The two types of customers are heterogeneous in service revenue and delay sensitivity, i.e., they have different revenue-cost structures. Besides, customers in class 1 can access to the delay information on servers' status and queue length as they arrive, but customers in class 2 can not. Distinguishing the two scenarios of whether there is a joining threshold for customers in class 1 in vacation state, we get the two types of customers' equilibrium joint joining strategies, respectively, and observe the interaction between the behaviors of the two types of customers. Moreover, we also get their socially optimal joint joining strategy for comparisons and for making control measures, and discuss the influence of the fraction of customers in class 1 on social welfare and system throughput. We find that in any scenario, the equilibrium social welfare is always unimodal with respect to the fraction of customers in class 1, which indicates that the higher fraction of customers in class 1 is not necessarily beneficial. However, on the premise of effectively controlling customer behavior, this high fraction is still necessary to optimize social welfare and increase system throughput, although the optimal social welfare is also unimodal.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call