Highly constrained college basketball scheduling

Abstract

In intercollegiate athletics, conference basketball schedules must satisfy several criteria to be acceptably fair to each team. Scheduling is often further complicated by building conflicts and agreements regarding televised games. Although these requirements make it tedious to develop feasible schedules by hand; they can readily be modelled as constraints in an integer programming formulation. Software has been developed that finds multiple feasible solutions to this integer program, where each solution corresponds to an acceptable schedule. By generating multiple schedules in a short period of time, the software allows conference personnel to quickly evaluate different scheduling scenarios (eg different TV games) and consider several alternatives. Thus, the software makes an important contribution to finding the most attractive schedule possible.