Meistaravörn við tækni- og verkfræðideild- Eva Linda Gunnarsdóttir
An Integer Programming Formulation for Scheduling of the Icelandic Football League
Friday May 31st kl. 13:00 Eva Linda Gunnarsdóttir will defend her 30 ECTS thesis in MSc Engineering Management. The title of the thesis is "An Integer Programming Formulation for Scheduling of the Icelandic Football League" and will be held in room M326. All welcome.
Student: Eva Linda Gunnarsdóttir
Supervisor: Eyjólfur Ingi Ásgeirsson, Associate Professor, Reykjavik University
Examiner: Páll Melsted, Professor, University of Iceland
Abstract:
In this thesis, the process of generating a sport schedule for a round robin competition is automated. By automating the process, the scheduling time should decrease and the schedule could potentially give a better solution by satisfying more requests from stakeholders.
The main goal for this project is to create a mathematical model by using integer programming and a two-phase approach. The presented model is used to make a schedule for the men’s highest division in the Icelandic football league in collaboration with the Icelandic Football Association (KSÍ). The current scheduling process is rather manual. KSÍ lean on a conventional table alignment from the regulations. The competition committee gathers team preferences and puts together a list of the main constraints. Next, they manually assign each team a number that best fulfills their wishes. The current process is time consuming and complex, this process can be greatly improved.
When generating the model, firstly a pattern set with equally distributed breaks is generated and used as an input for the two-phase mathematical model. In Phase I of the mathematical model a round robin schedule is generated while taking several hard constraints into account. A second model is generated to try and minimize the breaks in the pattern set. It is done using a large set containing all possible patterns. In Phase II of the mathematical model the teams are assigned to patterns using both hard and soft constraints and depending on team preferences. The schedule for the 2019 season had already been generated when this project was started. Therefore, preliminary results are presented by using real preferences and constraints sent in by the competition manager at KSÍ. These results are compared to the schedule which KSÍ generated.
The integer programming model is able to return a good solution while taking both fairness and break minimization into account, as well as satisfying most of the requests by using a two-phase model approach. After comparing the schedule from KSÍ to the preliminary results in this project it is apparent that the model could be of great help to the competition manager at KSÍ as well as a time saver.
