Allocating Games for the NHL Using Integer Programming

In recent years, the National Hockey League (NHL) expanded from 21 to 24 teams. In order to accommodate its scheduling process, the league had to determine new game allocations, that is, the number of times the teams play against each other. This paper introduces a procedure based on integer linear programming that generates several game allocation scenarios that have been presented to the league managers. The model takes into account the constraints specified by the NHL, and the objective function allows the league managers to specify their preference on the distribution of the games.