Order Quantity Determination Under Limited Demand Information

Abstract

This study shows the investigation of the maximum entropy approach for determining the order quantity of a single period newsvendor model under demand uncertainty. The characteristic of maximum entropy distribution is explored, i.e., forms of the estimated probability density function on different supports and availability of approximation methods. Comparisons among the solutions from the maximum entropy method, the solutions from the well-known distribution free approach and the optimal order quantities were conducted under the various demand distributions in the exponential family with identical means and variances. Additionally, this study extended the analysis to the situation where the higher order moment information is available.