2008 年 58 巻 6 号 p. 430-437
This paper proposes an optimal ordering method suitable for short lifecycle products whose demand is strongly dependent on the day of the week. Regarding application industries, we focus on retail stores such as $1 or convenience stores, most of which have limited space for display but a significantly large number of items. A considerable amount of research in inventory control has hitherto been carried out. We can, however, find few developments that can consider and handle the constraints imposed on the sector under consideration. Hence, this paper proposes a demand prediction and optimal ordering method utilizing the idea of a prediction limit that is common in regression analysis. The proposed method is advantageous in that it can be applied to newly introduced products, as it requires only a small number of samples for the parameter estimation. In the latter part of this paper, numerical simulations are performed using the actual sales records of a retail company, and the effectiveness of the proposed method is confirmed.