Many-server heavy-traffic limit for queues with time-varying parameters (Q2443196)

The paper under review is a sequel to [the authors, Queueing Syst. 71, No. 4, 405--444 (2012; Zbl 1275.60084)] that studied a deterministic fluid limit approximating \(G_t/GI/s_t+GI\) queues with general arrival process and time-varying arrival rate, independent and identically distributed (i.i.d.) service times, a time varying large number of servers, and customers abandonment from the queue with i.i.d. patience times. The paper under review establishes many server heavy traffic limits that provide mathematical support for both the previous fluid approximation in [the authors, loc. cit.] and refined Gaussian process approximation in the special case of exponential service time. On the basis of limit theorems, the paper under review proposes approximating the time varying number of customers in the system by a Gaussian distribution for each time \(t\). The model is assumed to alternate between underloaded intervals and overloaded intervals with critical loading only at the isolated switching points. The limits obtained in the paper under review extend the previous limits obtained for Markovian queues \(M_t/M/s_t+M\) in [\textit{A. Mandelbaum} et al., Queueing Syst. 30, No. 1--2, 149--201 (1998; Zbl 0911.90167); in: P. Key (ed.) and D. Smith (ed.), Proceedings of the 16th international teletraffic congress. Amsterdam: Elsevier. 355--364 (1999); in: Proceedings of the 37th annual Allerton conference on communication, control and computing. Allerton, IL: University of Illinois. 1095--1104 (1999); \textit{A. A. Puhalskii}, Math. Methods Oper. Res. 78, No. 1, 119--148 (2013; Zbl 1272.90010)].