Forecasting and Inventory Planning for Irregular Demand Patterns

Abstract

Medicine inventory in hospitals always faces either shortages or expired medication problems due to irregular demand. This work proposes three management approaches and compares their performances, including the Croston method, a combination of exponential and Poisson distribution (EPD), and the average inter-demand interval combined with average demand (AAD) were studied with the order-up-to-level policy. The exponentially weighted moving average (EWMA) control chart was used with some forecasting methods. Service level and average inventory were used to evaluate the performance of each approach. The study showed that the service level of each approach depended on the demand pattern characteristics. When the variance of demand and average demand were very high, the highest service level was found from the AAD with the EWMA approach. The Croston method was found more effective when the demand variability was relatively low, and demand appeared occasionally. Applying the EWMA increased the average inventory and service level when the AAD method was used.