Team Selection Using Multi-/Many-Objective Optimization with Integer Linear Programming
Shelvin Chand, Hemant Kumar Singh, Tapabrata Ray · 2018
Assembling a competitive team is a task encountered in many professional league sports such as cricket, soccer, rugby etc. Teams are assembled annually with players being bid for by competing franchises. While stochastic optimization approaches for team selection have been suggested in the past, the approximate nature of these techniques could be disadvantageous for team selection when the stakes are high. In this paper, we explore the use of multi-objective integer programming approach to alleviate this issue and deliver a set of optimal trade-off solutions (teams). We illustrate the performance of the approach using professional Twenty20 cricket league data from the Indian Premier League. We also demonstrate the ability to support partial team construction, i.e., selecting few members of the team with others unchanged. Lastly, we also present a way to rank the importance of the players within a team considering the key objectives.