Abstract
This paper presents certain theoretical and empirical results for problems which involve the assignment of priority ratings or indicators of military worth, where either several of these ratings must be aggregated or the outcome is to be evaluated by a multi‐valued criterion. The first portion of the paper contains a discussion of the logic of the problem and the axiomatic structure of a desirable solution. The second section includes the results of an experiment in which random demands were processed by running them through a model supply system using several alternative systems of priorities.