Otimização de equipes em League of Legends utilizando algoritmos genéticos multiobjetivo.

Abstract

League of Legends, a game of the Multiplayer Online Battle Arena or MOBA category (as it is popularly known), continues to be one of the highest paying electronic games in the world. This category is based on two teams that face each other on a symmetrical map with the objective of destroying the opposing base. One of the main points in this style of play and more specifically in League of Legends is the character selection stage (also known as champions), as it will guide the strategy of each team. In this step, players select which characters they will use within the game, where each character has characteristics and abilities that are different from the others. As it involves several factors in the selection process, it is considered a complex problem that can be solved with search techniques and artificial intelligence to find the best solutions. In this project, a new approach through Multi-Objective Evolutionary Algorithms (MOEA) is presented to generate teams in the game. In order to estimate the quality of the generated teams, a survey was conducted with a group of players. Significant results were achieved with this approach, obtaining an average rating of 4.5 for a total of 5 points.