Chris Pollett >
Students > [Bio] [Blog] |
Dou Di Zhu with Q-learningI use the python3 to create the Dou Di Zhu with Q-Learning In this part, I prepare three AI agents with separate Q-learning matrices These are my training and testing results:
1000 training times, 200 testing time result:
10000 training times, 2000 testing time result:
This simple Q-learning algorithm works not very well on Dou Di Zhu game, it only can teach these agents playing Dou Di Zhu like newbie players It's based on Python 3. Just run the 'main.py' to run this project. Deliverable 3.zip (You need to install the random library in your Python 3) |