Inventory control with unknown demand distribution : a slow-mover case

In this paper some tools will be developed for analysing a simple inventory problem of the slow-mover type with the special feature that the probability distribution of the demand interarrival times is unknown. The role of sufficient statistics in decision making is analysed for an arbitrary optimality criterion and particularly for the Bayesian criterion. A computational approach for the case of a Bayesian criterion is presented together with some numerical results.