Inventory control expert system for large scale retailers

Abstract

The needs of tightly coupled information system and inventory management system is caused by the progress of information system. Information system such as POS system or automatic warehousing system makes many assumptions and constraints needless, under which the existing inventory models have made an effort to resolve. Information system make it possible to use the complex but excellent neural network approach for demand forecasting and really neural network approach shows better performance than any other forecasting methods. By utilizing the information from demand forecasting process, we develop the (s*, S*) model with adaptive features based on the present (s, S) model. When making decision on safety stock and replenishment quantity, the (s*, S*) model determines adaptively one between two kinds of alternatives with respect to. Through the performance evaluation using real data, we prove that the (s*, S*) model is superior to the present (s, S) model in two measures of stockout occasion and inventory turnover. We suggest the architecture of the inventory control expert system to apply the (s*, S*) model to the real retailing industry and develop it. In order to validate the inventory control expert system in real industry, we apply it to Hanwha store case and the results are fairly good.