@inproceedings{8761053,
  abstract     = {{In a round-robin sports timetable, each team meets every other team a fixed number of times. The simplicity of this tournament structure notwithstanding, constructing a timetable is challenging due to the large number and diversity of constraints that typically need to be considered. A common approach is therefore to decompose the problem with the first-break-then-schedule approach (FBTS), which first determines when teams play home or away after which it decides upon the specific opponents. Despite common use of FBTS to schedule reallife tournaments in practice, research on how to backtrack between the different phases of the algorithms is scarce. This talk shows how FBTS and a closely related decomposition approach known as first-day-off-then-schedule (FOTS), relates to classic Benders’ decomposition with integer subproblems. Compared to the classic FBTS and FOTS approaches, traditional Benders’ decomposition provably converges to an optimal solution and is able to cope with most real-life constraints or even situations where the objective is (partly) determined by the assignment of opponents. In this talk, we generate compact timetables that minimize the total number of travel trips and relaxed timetables that minimize rest time penalties and differences in rest time of opposing teams. Despite the fact that various algorithms have been proposed in the literature before, we find considerably better solutions for several instances of all three applications.}},
  author       = {{Van Bulck, David and Goossens, Dries}},
  booktitle    = {{EURO 2022 : conference handbook and abstracts : 32nd European Conference on Operational Research (EURO XXXII)}},
  isbn         = {{9789519525419}},
  language     = {{eng}},
  location     = {{Helsinki, Finland}},
  pages        = {{65--65}},
  title        = {{Backtracking in sports timetabling : a traditional Benders’ decomposition approach}},
  url          = {{https://www.euro-online.org/conf/admin/tmp/program-euro32.pdf}},
  year         = {{2022}},
}