An Approach for Team Composition in League of Legends using Genetic Algorithm

In recent years, Multiplayer Online Battle Arena and League of Legends are the most played genre and game. This is a genre which competitors are separated into two teams facing each other in a common objective, that generally involves destroying the opponent base. One of the most important steps in LoL is the selection of the champions who will be used in the match, since this involves many factors and variables that are important to the match such as attributes and abilities of each character. Due to several variables, team composition is considered a complex problem and can be handled with search-based algorithms. This paper proposes an automated team composition approach in League of Legends by combining attributes of the champions and game strategies via a Genetic Algorithm. It aims to generate teams from total set options in the game by focusing on satisfying constraints contained in real game strategies. Thus, this study also presents a fitness function to assess the adequacy of the generated teams. Finally, the paper reports empirical results regarding the effectiveness and efficiency of the proposed approach on three game strategies. Our results demonstrate that the proposed approach is useful for this context since the generated teams achieve high adequacy regarding the objective functions in a low time.