Designing adaptive opponent models for RTS games

Games, reality of nature, have a very important part in our life. Games are fed by challenges and real challenges can only be created with help of AI. This thesis deals with corner kicking in soccer game. In thesis, it is studied in RoboCup 2D simulation league with Dainamite team against Wright Eagle team and focused on corner kicking from a different aspect than other all RoboCup teams. While corner kick situations are all the time treated by an ordinary short pass or kick in situation by other teams, in this thesis it is treated as corner kick that allows carom positions or running players which requires a long kick inside the penalty area. In corner kick, it is only decided where to kick the ball, positioning of players in the field is excluded. For the classification method, Support Vector Machine?s LibSVM implementation is used and experiments are performed on parameters of SVM and MySVM implementation.