Random-weight Generation in Multiple Criteria Decision Models

In this paper, we consider generating random uniformly-distributed weights for the simulation of additive p-criteria decision models. First we review several approaches for generating random weights, and show that some obvious approaches do not work well at all. Then, we consider several procedures for generating weights that satisfy general linear constraints. Instead of solving many small linear programming problems to find pairwise dominance of alternatives, we may solve such problems using simulation.