Abstract
The control and management of inventory is becoming more and more important for enterprises, especially in today’s competitive marketplace. An effective inventory control can provide huge competitive advantages to a company. This study considers a dynamic single-item inventory problem with uncertain demands in a rolling horizon environment. In this problem, decision maker has to determine how and when to make fulfillment orders for future demands with the objective of minimizing total cost. Such an inventory control problem has to make the trade-off decision between ordering, holding, and backorder costs. There are two decision models presented in this study for the investigated inventory control problem. One is a deterministic model, in which the problem information like various costs and future demands are known constant parameters. The other one is a stochastic model, in which some of the problem information is uncertain. The uncertainty of future demands is presented by a number of scenarios; the chance of each scenario is represented by a probability. The stochastic model in this study is a scenario-based mixed integer programming (SBMIP). Given constant know future demands, a dynamic programming (DP) is used to solve the deterministic inventory control problem. The focus of this study is to compare the long-term performances of these two models under rolling horizon environment via simulation experiments. In the experiments, a paired -test is used to compare the performances of these two decision models under different control factors. From the results of simulation experiments, we observe that the performance of the deterministic model is superior to the stochastic one in the dynamic single-item inventory problem with uncertain demands in a rolling horizon environment.