Forecasting and inventory area model choice

Abstract

Inventory control is an important element of both business and military cost control and readiness. The Taiwan Army Logistic Agency (TALA) has used a combination of mathematical inventory models, arithmetic average, three month moving average and experience to project future demand. Implicit is that the mean of monthly demand for an item remains steady over time. This assumption has proven to be incorrect during periods of force reduction, equipment is replaced or retired, or when there is a cyclical demand. Once an unusual demand pattern occurs, inventory control becomes unpredictable. Inapplicable inventory methods in the TALA have been estimated to cost as much as several million dollars. TALA has focussed on advanced forecasting methods, Exponential Weighted Moving Average (EWMA)to solve this problem. This may reduce inapplicable inventory to some extent. Residual inventory and shortage are also factors in cost control. In this research we will explore the appropriate approach to solve these problems to make the inventory control more efficient.