A finite horizon inventory model: An operational framework

We present a simulation based decision support system motivated by a real life situation to decide the inventory ordering policy in the context of a single commodity, multipack, and finite horizon situation. The multiple objectives include (a) minimizing the end of the season inventory, (b) maximizing the operating profit, and (c) minimizing the peak working capital requirements during the season. Stochastic demand and positive lead time add to the complexity of the problem context. In addition, multiple partners in the supply chain with distinct and conflicting set of objectives necessitate the need for a formal approach. The problem can be formulated as a multiperiod dynamic programing problem with stochastic demand with an objective to optimize the expected operating profit, subject to specific constraints on working capital requirement, service level, order fill rate, and end of the season inventory. Such a formulation is hard to solve analytically and does not lend itself to analyze several ordering policies (what-if analysis). Based on simulation experiments, we propose an ordering policy which optimizes the overall objective(s) of supply chain partners. The model is simple and easy to use and is designed with adequate flexibility to conduct what-if analysis.