Game AI

Strategy selection and swarm intelligence for real-time strategy games, and RoboCup rescue simulation.

Real-time strategy games such as StarCraft are a hard testbed for AI: many units, partial information and decisions in real time. With colleagues at UFMG we worked on evolving swarm intelligence for task allocation (Best Paper Award, SBGames 2014) and on game-theoretic strategy selection (AIIDE 2016). Our bots placed 3rd in the IEEE CIG 2016 student StarCraft competition and in the top 10 of the AIIDE 2016 competition, and our team took 2nd place in the Rescue Agent Simulation league of the Latin American RoboCup 2015.

Code on GitHub

Publications

  1. JIS
    Combining genetic algorithm and swarm intelligence for task allocation in a real time strategy game
    Tavares, Anderson R., Zuin, Gianlucca Lodron, Azpúrua, Héctor and Chaimowicz, Luiz
    Journal on Interactive Systems 2017
  1. AIIDE
    Rock, Paper, StarCraft: Strategy Selection in Real-Time Strategy Games
    Tavares, Anderson, Azpúrua, Héctor, Santos, Amanda and Chaimowicz, Luiz
    In Artificial Intelligence and Interactive Digital Entertainment 2016
  2. RoboCup
    Y-Rescue Team Description Paper
    Tavares, Anderson, Saldana, David, Azpúrua, Héctor, Pimentel, Jhielson, Vidal Pino, Omar and 3 more authors Tavares, Anderson, Saldana, David, Azpúrua, Héctor, Pimentel, Jhielson, Vidal Pino, Omar, Colares, Rafael, Macharet, Douglas and Chaimowicz, Luiz
    In RoboCup 2015
  3. SBGames
    Evolving Swarm Intelligence for Task Allocation in a Real Time Strategy Game
    Tavares, Anderson R., Azpúrua, Héctor and Chaimowicz, Luiz
    In 2014 Brazilian Symposium on Computer Games and Digital Entertainment (SBGAMES)
    Best Paper Award, SBGames 2014