Lot Sizing using Neural Network Approach

Abstract

A lot of works have been done by the researchers to solve lot-sizing problems over the past few decades. Many techniques and al-gorithm have been developed to solve the lot-sizing problems. Basically, most of the algorithms are developed either based on heuristic or math-ematical approach. Since neural network has been given attention by the researchers in many areas including production planning, therefore in this paper we implement neural network to solve single level lot-sizing problem. Three models are developed based on three well known heuris-tic techniques, which are Periodic Order Quantity (POQ), Lot-For-Lot (LFL) and Silver-Meal (SM). The planning period involves in the model is 12 period where demand in the periods are varies but deterministic. The model was developed using MatLab software. Back-propagation learning algorithm and feed-forward multi-layered architecture is cho-sen in this project. Result shows that the three models able to give optimum solution and easy to be applied in the lot-sizing problem.