In this paper we propose a state-dependent importance sampling heuristic to estimate the probability of population overflow in networks of parallel queues. This heuristic approximates the “optimal��? state-dependent change of measure without the need for costly optimization involved in other recently proposed adaptive algorithms. Preliminary results from simulations of networks with up to 4 parallel queues and different traffic intensities yield asymptotically efficient estimates (with relative error increasing sublinearly in the overflow level) where state-independent importance sampling is ineffective.
|Title of host publication||The 2006 Russian-Scandinavian Symposium on Probability Theory and Applied Probability|
|Number of pages||3|
|Publication status||Published - 2006|
Zaburnenko, T. S., & Nicola, V. F. (2006). Efficient Heuristics for Simulating Population Overflow in Parallel Networks. In The 2006 Russian-Scandinavian Symposium on Probability Theory and Applied Probability (pp. 86-93)