Assignment for AI Planning for Autonomy (COMP90054). Various algorithms are analysed and thoroughly evaluated to design an autonomous player for the modified board game Sequence.
The python version supported is python 3.6+.
All required classes and methods to implement a new game are defined in the template.py
Only additional import required is the func_timeout module. Install with: pip3 install func_timeout
There are three give general agents that will work with any game under directory agents/group76: heuristic.py, qLearning.py and deepQ.py. In order to guide the AI's decision making, the methods implement a heuristic function, Q-Learning and Deep Q-Learning respectively.
The game can be run with specified runner. The only difference between runner is the first two line of the code (importing different game files.) The options can be found using following command:
python sequence_runner.py -h
If running Sequence, note that the game will start in fullscreen mode. Press F11 to toggle fullscreen. The game's activity log now appears as a separate window.
- save the print as log
- save replay
- play saved replay file
- run multiple games in sequential
- Sequence
For teams to import their customized python file, they will have to add their team name into the path. For example:
import agents.staff_team_random.timeout