Automated Timetable Generation for Rounds of a Table-Tennis League

In this research we consider the problemof scheduling rounds of a non-professional table-tennisleague. We formalize the problem in terms of atimetabling optimization problem. Then we solve thishighly constrained problem with a permutation based GeneticAlgorithm for which feasibility preserving operatorsare defined. Since coding and operators cannot warrantfeasibility in every case, the fitness function penalizes constraintviolations. This algorithm is compared to an evenmore elaborated variant, which additionally aims at repairinginfeasible solutions produced by the genetic operators.