|
Abstract : |
Computer simulation is an important tool in the analysis and design of communications networks. In spite of the advances in computational power, using simulation to obtain rare event probabilities such as cell/packet loss or delay in networks still requires prohibitively long execution times. We provide an overview of importance sampling techniques and how they can be used to provide orders of magnitude speedup for many network problems. of the target rare event(s). In this sense, any important measures of performance in com-M munications networks are defined in terms of rare trajectory splitting-based techniques are clearly a type of IS. Trajectory splitting achieves speedup by launching additional subtrajectories from intermediate system event probabilities. Two examples are cell loss probability and states. The intermediate states are visited much more often cell delay threshold probability in asynchronous transfer mode than the target states themselves and behave as gateway states, |
