Search

How would the growing prevalence of real-time delay information affect a service system? We consider a single-server queueing system where customers arrive according to a Poisson process and the service time follows an exponential distribution. There are two streams of customers, one informed about real-time delay and the other uninformed. The customers' uninformed behavior may be due to information ignorance or rational behavior in the presence of an information fee. We characterize the equilibrium behavior of customers with information heterogeneity and investigate how the presence of a larger fraction of informed customers affects the system performance measures, i.e., throughput and social welfare. We show that the effects of growing information prevalence on system performance measures are determined by the equilibrium joining behavior of uninformed customers. Perhaps surprisingly, we find that throughput and social welfare can be unimodal in the fraction of informed customers. In other words, some amount of information heterogeneity in the population can lead to more efficient outcomes, in terms of the system throughput or social welfare, than information homogeneity. For example, under a very mild condition, throughput in a system with an offered load of 1 will always suffer if there are more than 58% of informed customers in the population. Moreover, it is shown that for an overloaded system with offered load sufficiently higher than 1, social welfare always reaches its maximum when some fraction of customers is uninformed of the congestion level in real time. [ABSTRACT FROM AUTHOR]

This paper provides the first exact analysis of a preemptive M=M=c queue with two priority classes having different service rates. To perform our analysis, we introduce a new technique to reduce the two-dimensionally infinite Markov chain (MC), representing the two class state space, into a one-dimensionally infinite MC, from which the generating function (GF) of the number of low-priority jobs can be derived in closed form. (The high-priority jobs form a simple M/M/c system and are thus easy to solve.) We demonstrate our methodology for the c = 1,2 cases; when c >2, the closed-form expression of the GF becomes cumbersome. We thus develop an exact algorithm to calculate the moments of the number of low-priority jobs for any c ≥ 2. Numerical examples demonstrate the accuracy of our algorithm and generate insights on (i) the relative effect of improving the service rate of either priority class on the mean sojourn time of low-priority jobs; (ii) the performance of a system having many slow servers compared with one having fewer fast servers; and (iii) the validity of the square root staffing rule in maintaining a fixed service level for the low-priority class. Finally, we demonstrate the potential of our methodology to solve other problems using the M/M/c queue with two priority classes, where the high-priority class is completely impatient. [ABSTRACT FROM AUTHOR]

The most common measure of waiting time is the overall expected waiting time for service. However, in service networks the perception of waiting may also depend on how it is distributed among different stations. Therefore, reducing the probability of a long wait at any station may be important in improving customers' perception of service quality. In a single-station queue it is known that the policy that minimizes the waiting time and the probability of long waits is nonidling. However, this is not necessarily the case for queueing networks with several stations. We present a family of threshold-based policies (TBPs) that strategically idle some stations. We demonstrate the advantage of strategically idling by applying TBP in a network with two single-server queues in tandem. We provide closed form results for the special case where the first station has infinite capacity and develop efficient algorithms when this is not the case. We compare TBPs with the nonidling and Kanban policies, and we discuss when a TBP is advantageous. Using simulation, we demonstrate that the analytical insights for the two-station case hold for a three-station serial queue as well. [ABSTRACT FROM AUTHOR]