Customer Threshold Strategies in Observable Queues with Partial Information of Service Time

Abstract

This paper concerns customers’ decision process of joining or balking at their arrival instants in an observable queue. Given partial information that different number of moments of the service time is available to the arriving customers, they need to estimate its distribution by using only the information available and no other information based on the maximum entropy principle to decide whether join or balk. Respectively for the systems with risk-neutral and risk-averse customers, we numerically find that in general the equilibrium thresholds are no less than the socially optimal one which is also no less than the profit-maximizing one.