Dynamic Modeling of Inventories Subject to Obsolescence

A class of models for optimizing inventory costs is presented which takes account of stochastic obsolescence of an inventory item. Obsolescence is defined as a demand state, in such a fashion as to permit appraisal, ex ante, of the probability of arrival of obsolescence at future times, under the assumption that there are many possible states of demand. Response to obsolescence is introduced by means of a Bayesian procedure. In the most complete model this is done by modification of the state probability vector of a Markov process. Optimization is accomplished by means of a dynamic program.