Simpósio Latino-Americano de Jogos

Utilização de rede neural artificial otimizada no aprendizado de máquina em jogos digitais

Renan Cunha, Luan C. Casagrande, Yuri Crotti, Roderval Marcelino, Vilson Gruber

Resumo

In recent years, a lot of research has been done about the use of artificial intelligence in games. It has been proven that neural network, for example, can be used in a huge number of possibilities in games. Considering this fact, the authors proposed a novel methodology for training neural networks as airplane controller. Our effort is focused to optimize the result using genetic algorithm to find the best topology and parameters for our problem. The proposed system was modeled, tested and validated under a simulation platform using flight simulator FlightGear.