The prediction of the future states in Multi Agent Systems such as Robocup Soccer has been a challenging problem since the beginning of the MAS. Robocup 3D Soccer is selected because of its global view of the agents. An Artificial Neural Network is used for prediction. The goal is to concentrate on the design of an optimal ANN. In order to handle it, a genetic algorithm is used for optimization of the design, which shows a great improvement over the manual methods.
M. Moghimi Najaf Abadi, C. Lucas
University of Tehran