F-025
High season demand forecasting method in logistics warehouse inventory management

◎陳　　成・尾白大知・平山淳一・荒　宏視（日立）

Demand forecasting is a core of logistics warehouse inventory management. A challenge in logistics demand forecasting is high season (peak) demand forecasting. High season with huge number of demand is hard to predict because of data limitation and shift of high season. This paper proposes a high season demand forecasting method extracting high season of each item by combining single item and all items’ high season probability distribution, and predict high season’s demand by a special model different to normal season. The experiments show proposed method decreases demand prediction error amount including high season 7.8% compare to existing methods.