On Some Regenerative Approaches of Queuing Network Simulation

The aim of this work is to develop regenerative approaches based on discrete time events simulation and framework of general semi-markov processes (GSMP). It’s expanding field of classic methods for research of queuing networks, which are simulated ineffective (or impossible at all) with classical regeneration. We suggest that multidimensional markov process describe system. It consist of discrete (for example, vector of number of customers in a network) and ???-continuous components. Continuous components mean time counters (clocks) remained time until next discrete components transformation. Discrete components change at moments of occurring some events (for example, arrival/depature customers). Also, at such moments, recalculating of several (or all) continuous components proceeded.