HOW DISCRETIZATION AFFECTS INTERMITTENT DEMAND STOCK MANAGEMENT BASED ON SIMULATION

Abstract

This paper is aimed at the development of an alternative combinatorial strategy of reducing searched solution space in intermittent demand stock management based on the past stock movement simulation. The combinatorial strategy involves an adjustable level of the discretization of control variables that are used within a selected inventory control policy. We combine this new strategy with the local search employing linear regression and bootstrapping to bound the reorder point and simulate (Q, R) inventory control policy using randomly generated data. The data is characteristic with an increasing intermittency and a non-zero demand variability. The outputs from simulation experiments show that combining two different strategies of reducing searched solution space brings a significant improvement in the trade-off among the minimal holding and ordering costs, required service level and the consumption of the computational time making the past stock movement simulation to be more applicable in extensive real life tasks.