Optimal Ordering Policies for Stochastic Inventory Problems with Observed Information Delays

Abstract

Information delays exist in an inventory system when it takes time to collect, process, validate, and transmit inventory/demand data. A general framework is developed in this paper to describe information flows in an inventory system with information delays. We characterize the suffcient statistics for making optimal decisions. When the ordering cost is linear, the optimality of a state-dependent base-stock policy is established even when information flows are allowed to cross over time. Additional insights into the problem are obtained via a comparison between our models and the models with stochastic order lead times. We also show that inventory can substitute for information and vice versa.

Keywords: Information delay, partial observations, stochastic inventory problem, incomplete information, base-stock policy, dynamic programming, sufficient statistic, destination and origin determined dynamic delays, Information flow crossing

JEL Classification: C61, M11, D81, D83