The impact of information about last customer’s decision on the join-or-balk dilemma in a queueing system

Strategic customer behavior regarding the join-or-balk dilemma in queueing systems has been studied intensively under various kinds of information structures. The majority of these studies focus on the observable and the unobservable cases, where an arriving customer observes or does not observe, respectively, the number of present customers before making her decision. An important finding is that more information does not always improve customers’ and/or the administrator’s benefits and may result to a deterioration of a system. Therefore, intermediate information structures have been proposed that bridge the two extreme cases: partially observable models, models with delayed observations, alternating observable models etc. All these structures revolve around the idea that the administrator of a service system should control somehow the information about the state of the system, which is usually the number of present customers. In this paper we consider a new mechanism which consists in informing customers about other customers’ decisions. Such a mechanism helps customers to coordinate themselves and possibly leads to better outcomes. To present this idea in the simplest possible framework we consider the M/M/1 queue with strategic customers that face the join-or-balk dilemma and assume that each arriving customer is informed about the decision of the most recent arrival. We show that this system outperforms the observable and unobservable systems for certain ranges of the parameters. Moreover, the effective arrival process is more regular, a fact that improves several performance measures of the system.