DECISIVE ARRIVAL LAW PARAMETERS AND A GENERAL FINITE-CAPACITY QUEUINGPROBLEM

This paper addresses the problem of finding the parameters of the arri val law which most significantly influence expected occupation and los s of a finite capacity queue. The input process is supposed to be ergo dic and wide sense stationary. We show that it is mostly possible to f it an MMPP(2) to the decisive parameters of observational data. Numeri cal examples illustrate the importance of the decisive parameters, cal led key parameters, and also show the accuracy of the proposed fitting procedure. Finally, in the appendix we present the solution of the fi nite capacity queueing problem with Special Semi Markov Process (SSMP) arrivals and a general service strategy.