Benchmarking of a stochastic production planning model in a simulation testbed

Abstract

A major problem in production planning is to determine when to release products into production to meet forecasted requirements. Riano et al. (2002) proposed the stochastic production planning (SPP) model for a multiperiod, multiproduct system, where the lead time to produce a product may be random. The model determines release times for the products that ensure the requirements in each time period are met with desired probabilities at a minimum cost. We describe how an advanced planning model like SPP can be integrated with discrete event simulation models to make the simulations more realistic and informative. We also compare the performance of the SPP model with the classical MRP (materials requirements planning) model, and with a stochastic variation of the MRP model in a simulation study. The costs associated with the production plans from SPP are about 10% less than the costs from the other two models.