A Research on Non-Stationary Stochastic Demand Inventory Systems
Dr. Mehmet Soysal

Abstract
Non-stationary stochastic lot sizing problem appears in many industrial environments where demand is nonstationary and stochastic, and replenishments require fixed freight fees. This paper presents the cost performance assessment of three inventory control strategies proposed for the stochastic lot sizing problem, namely Hua et al. (2008)’s strategy (HYHX), Tarim and Kingsman (2006)’s model (TK) and rolling horizon integrated version of Tarim and Kingsman’s model (Rolled TK), against the optimal policy, non-stationary (s, S). The computational analyses show that the rolling horizon method implementation to TK escalates the model’s performance, and HYHX and Rolled TK perform quite close to each other in terms of expected cost.

Full Text: PDF