Optimization of the structure of the speedway team in the presence of regulation and financial constraints –application of machine learning and integer programming to the Polish league case

Abstract

In a difficult macroeconomic environment, managing a sport club and a sport team is not an easy task. It is a special challenge due to financial, demographic and organizational constraints arising and changing permanently. One of the main tasks faced by managers in the team sports is to build the optimal composition of the team, taking into account the measures that define the sports goal. The article discusses the problem of team structure optimization with the use of statistical data on the example of the Polish highest level speedway league, PGE Ekstraliga. Taking into account the demographic predictor (age of the riders) and the nationality of the riders, an analysis of the importance and impact of these predictors on the rider's performance was carried out. Using machine learning methods, the characteristic profiles of the riders, differing in effectiveness, were also indicated. The method used allows for local interpretation of prediction, as well. As a final step of the research an optimization mechanisms of team building was developed. The study proposes an optimization model supporting the decision making process of creation of the structure of the team registered for the meeting. Integer programming optimization mechanisms were constructed in concordance with possible financial and organizational constraints which could be faced by team managers. Models’ validation with the use of numerical parameters’ values showed their internal consistency and provided interesting decision’s indications. Both machine learning method and optimization models can be develop further to cover more reach data sets and more managerial constraints or even could be tuned to be applied in the other team sports.