Sport Tournament Scheduling with Constraint Programming

In double round robin sport competitions, each team plays each other team twice, once at home and once at the other team's place. The coordinators of such tournaments often have to deal with a host of requirements from the participating teams, the fans and the media. The result is a combinatorial search problem of considerable complexity.

Our research suggests that constraint programming allows to solve such tournament planning problems. We could show that a constraint programming solution to a difficult example, the ACC 1997/98 problem, is orders of magnitude faster than previously existing approaches.

Based on this experience, we are building the tool Friar Tuck, a generic double round robin tournament planner that allows to conveniently enter a variety of constraints. Friar Tuck will allow the coordinator of (especially professional) sport tournaments to compute provably optimal solutions to his or her tournament planning problems. Friar Tuck is being implemented in the concurrent constraint language Oz, using the programming system DFKI Oz 2.0.

This work is currently funded through the project ReAlloc.