On switching-type mixed-ordering policy based on monitoring forecasting error

Abstract

In a fixed-size ordering system we order a constant economical lot size whenever current inventory breaks re-order point. On the other hand, in a fixed-interval ordering system we periodically order the required quantity by taking into account forecasted demand and term-end inventory. In this paper, we show a theoretical criterion of choice between these two policies, and then propose a mixed-ordering policy which builds in elements of both types of ordering systems. The theoretical criterion tells us that a fixed-interval ordering policy becomes preferable when we make forecasting error small, but otherwise, a fixed-size ordering policy is better. Therefore, the mixed optimal policy is designed to dynamically switch between these types of ordering systems by monitoring forecasting error. We shall prove the existence of an optimal mixed strategy.