SUMMARY
Key to the problem of stock management is the uncertainty of product demand. What is the best way to include the consideration of demand uncertainty in models that are both effective and applicable to the day-to-day complexities facing managers? This paper presents two innovative ways of considering demand uncertainty in stock control. The first approaches uncertainty as a stochastic phenomenon, which we address by using a Stochastic Dynamic Programming (SDP) model. The second approaches demand uncertainty as a fuzzy-type phenomenon, which we address using Fuzzy Inference System(FIS). Our focus is to explore the application of Fuzzy Inference System to a classic problem of stock control: the definition of order sizes. For a given demand scenario, we evaluate the two approaches by calculating their total cost’s present value and by comparing their cost performance. The cost evaluation method used was a Two-Dimensional Fuzzy Monte Carlo Simulation, in which 10,540 demand scenarios were analyzed. Our results show that the SDP model results in lower total costs compared to the FIS model. We also found out that the distributions of FIS decisions were similar to the SDP and that costs varied up to 40% from the optimal value. Thus, in a situation where it is either impossible or unfeasible to solve the problem using SDP, Fuzzy Inference Systems can be utilized as a relatively simple and effective alternative.