SCSC2003 Abstract S21122
An Efficient Sampling Technique for Stochastic Supply Chain Simulations
An Efficient Sampling Technique for Stochastic Supply Chain Simulations
Submitting Author: Ms. Wing Yan Hung
Abstract:
To enable a fair comparison of various supply chain initiatives, realistic supply chain simulation models need to capture the system dynamics and characteristics of individual supply chain members by modelling their physical and business activities. Of equal importance is the need to quantify the supply chain performance under uncertainty. This requires detailed stochastic models that are usually very high in dimension and thus require extensive computation. In this paper we present an efficient sampling technique to reduce the number of simulations required to obtain accurate performance metrics. Latin Supercube Sampling (LSS) based on Sobol’ subsets is applied to generate the sample points for the uncertain variables. Numerical experiments and a case study are used to illustrate the higher efficiency of this sampling method.
Back to SCSC2003 Abstracts