REGENERATIVE RARE EVENTS SIMULATION VIA LIKELIHOOD RATIOS

In this paper we obtain some new theoretical and numerial results on e stimation of small steady-state probabilities in regenerative queueing models by using the likelihood ratio (score function) method, which i s based on a change of the probability measure. For simple GI/G/1 queu es, this amounts to simulating the regenerative cycles by a suitable c hange of the interarrival and service time distribution, typically cor responding to a reference traffic intensity rho0 which is < 1 but larg er than the given one rho. For the M/M/1 queue, the resulting gain of efficiency is calculated explicitly and shown to be considerable. Simu lation results are presented indicating that similar conclusion hold f or gradient estimates and in more general queueing models like queuein g networks.