TLSG.mp4

Описание: The aim of this project is to develop agents which have ability to learn about their opponent and use what they learn to improve tactics to oppose the opponents. The project was simulated in form of RTS game style which include two different sides of agents consisting of the defender and attacker. Q learning is used to control the defender to learn from the attacker. The attacker side is controlled by FSM (Finite State Machine) and Neural Network. However, it is optional for the attacker to be controlled by real player as well. C++, OpenGL and OpenAL were used to complete this project. Currently, I am working on Quadtree, Path finding and also more funtionallities in the game. Any update of this project will be posted here. Please feel free to comment on it. Thank you for taking your time to watch it.