How victor-lee-123/ai-q-learning-project works
This Unity project is a Q-learning game demo, documented as a practical reinforcement-learning example and distributed with a runnable ChasingArena build. The apparent user workflow is to launch the arena, observe or interact with an agent, and see training or game information; the README points to a PDF for demo details, but that document and the runtime source were not excerpted. Components below are identified from the project tree and filenames. Dashed, uncited links are tentative, not verified call paths.
Main entry points: SampleScene.unity Arena scene · TrainingManager.cs Training manager · AgentController.cs Learning agent · Agent.prefab Agent prefab
Components 10
Arena gameplay
- Learning agent
SourceCode/Assets/Scripts/AgentController.cs - Agent state
SourceCode/Assets/Scripts/AgentInfo.cs - Arena scene
SourceCode/Assets/Scenes/SampleScene.unity - Agent prefab
SourceCode/Assets/Prefabs/Agent.prefab - Floor prefab
SourceCode/Assets/Prefabs/Floor.prefab
Q-learning
- Q-table
SourceCode/Assets/Scripts/QTable.cs - Training manager
SourceCode/Assets/Scripts/TrainingManager.cs
Input and UI
- Player controller
SourceCode/Assets/Scripts/PlayerController.cs - Game UI
SourceCode/Assets/Scripts/UIManager.cs
Other components
- Player
Connections 9
- Player → Player controller: provides input
- Player controller → Arena scene: interacts with
- Arena scene → Agent prefab: uses
- Arena scene → Floor prefab: uses
- Agent prefab → Learning agent: hosts
- Learning agent → Agent state: uses state
- Training manager → Learning agent: trains
- Training manager → Q-table: updates values
- Game UI → Training manager: displays status