GOAL-SEEKING PROBLEM IN DISCRETE-EVENT SYSTEMS SIMULATION

For most complex stochastic systems such as microelectronic systems, t he Mean Time To Failure (MTTF) is not available in analytical form. We resort to Monte-Carlo Simulation (MCS) to estimate MTTF function for some specific values of underlying density function parameter(s). MCS models, although simpler than a real-world system, are still a very co mplex way of relating input parameter(s) to MTTF. This study develops a polynomial model to be used as an auxiliary to a MCS model. The esti mated metamodel is a Taylor expansion of the MTTF function in the neig hborhood of the nominal value for the parameter(s). The Score Function methods estimate the needed sensitivities (i.e. gradient, Hessian, et c.) of the MTTF function with respect to the input parameter in a sing le simulation run. The explicit use of this metamodel is the target-se tting problem in Taguchi's product design concept: given a desired tar get MTTF value, find the input parameter(s). A stochastic approximatio n algorithm of the Robbins-Monro type uses a linear metamodel to estim ate the necessary controllable input parameter within a desired accura cy. The potential effectiveness is demonstrated by simulating a reliab ility system with a known analytical solution. Copyright (C) 1996 Else vier Science Ltd.