Information and source code for the RLTankAttack game - Developed as part of Master Thesis
Machine Learning Mechanisms and their Use in Validating the Quality of Novel Games
Aug 31, 2012
Open University of Cyprus
Our objective is to use existing knowhow with reinforcement learning and its implementation technology to automatically discover tactics in a game, so that the learning mechanisms can be used in novel games, currently uder developement. In this thesis we are trying to design a new zero-sum between two players and validate its properties (e.g fairness based on a random game, learning speed based on its complexity etc.) and the computers ability to learn by observing other players (humans or not).