IMPORTANCE SAMPLING SIMULATION IN ULTRASAN

Traditional simulation techniques perform poorly when estimating perfo rmance measures based on rare events. One solution to this problem is the use of importance sampling. However, two problems that have limite d the use of importance sampling are the lack of a formal framework fo r specifying importance sampling strategies, and the fact that in most cases the simulations must be hand-coded - a very time-consuming proc ess. This paper presents a software tool that facilitates experimentat ion with importance sampling by addressing the two problems. First, th e tool is based on a flexible framework for specifying importance samp ling simulations in terms of stochastic activity networks. Second, onc e specified, the importance sampling simulation program is automatical ly generated by the tool, freeing the researcher to focus on the model ing problem. The effectiveness of the software is demonstrated through the solution of a machine-repairman model with Weibull distributed fa ilure times and a delayed group repair policy. Orders of magnitude red uction in the CPU time required to obtain a specified relative accurac y were achieved.