Lot Sizing of Multistage Manufacturing System under Fixed and Rolling Environments

This paper proposes a Neural network (ANN) as a heuristic for multistage lot sizing. The objective of inventory management is to maintain optimum levels of inventory consistent with customer demands and plant capacity. Stated simply, management must determine what to order, when to order, and how much to order. This is not an easy task, for there are many conflicting goals. Nevertheless, management must inevitably make a decision to order what it considers to be an “economic lot size”. In this paper the influence of various parameters of lot sizing of multistage manufacturing system under fixed and rolling environments is detailed. Design of experiments methodology is used in this onnection. In the rolling horizon, the behavior of the neural network and other rules are compared with and without freezing the plan. The performance of the neural network is compared with the Wagner Whit in algorithm and part period algorithm. The superiority of the proposed method is discussed.

Keywords:

Neural network , Lot Sizing , Design of Experiments , Fixed Environment , Rolling Environment