Coordinated Replenishment Game and Learning Under Time Dependency and Uncertainty of the Parameters

Abstract

This research proposes a periodic review multi-item two-layer inventory model. The main contribution is a novel approach to determine the can-order threshold in a two-layer model under time-dependent and uncertain demand and setup costs. The first layer consists of a learning mechanism to forecast demand and forecast setup costs. The second layer involves the coordinated replenishment of items, which is analysed as a Bayesian game with uncertain prior probability distribution. The research builds on the concept of the (S, c, s) policy, which is extended to the case of uncertain and time-dependent parameters.