Optimal realignment of athletic conferences

Abstract

This article presents a mathematical analysis of the current realignment plan for men’s, Division I, college hockey. Comparisons are made between existing alignments, proposed alignments, random alignments, and optimal alignments, with respect to various measures. It is shown how proposed alignments do not minimize travel distance nor maximize attendance. It is shown also how a non deterministic clustering procedure can be expected to outperform the proposed alignments, with respect to these measures. Although this clustering procedure is not almost surely optimal, it is shown in our hockey setting to be an effective approximation, being nearly optimal and easily computable. R programs are provided in the appendix.