A Four-Moments Alternative to Simulation for a Class of Stochastic Management Models

This paper presents a computational alternative to simulation for a large class of stochastic management models involving functions of random variables. An example of a model in this class is the well-known “risk analysis” problem studied by Hertz and Hillier. Our computational approach includes (i) a versatile framework to describe the univariate and dependence characteristics of a model's random variables, and (ii) formulas for computing the central moments of the model's objective variable. The usefulness of these central moments in decision making is then illustrated and discussed.