This game was developed as a coursework for the AI module of my MSc in Computer Games Technology. The project is a real strategy game which features an arena filled with buildings and units, which was carefully designed in order to safely train our AI in parallel instances at the same time.
The game features two different AIs implementations:
- Rule-based system
- Reinforcement Learning via MlAgents
The following screenshot displays the Ray Perception Sensor 3D in use by the AI during the reinforcement learning training and execution.
Another screenshot of the game from above:
