GitHub - sicotronic/q-tetris_ia: IA agent that learns through reinforcement learning to play tetris as best as posible, implemented with the Q-Learning algorithm

sicotronic / q-tetris_ia Public

Notifications You must be signed in to change notification settings
Fork 0
Star 2

IA agent that learns through reinforcement learning to play tetris as best as posible, implemented with the Q-Learning algorithm

GPL-2.0 license

2 stars 0 forks Branches Tags Activity

Star

Notifications

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
agents/tetrisAgentJava		agents/tetrisAgentJava
proving		proving
system		system
trainers		trainers
COPYING		COPYING
Makefile		Makefile
README		README
install.bash		install.bash
runDemo.bash		runDemo.bash

Repository files navigation

This is a java IA agent that is meant to select actions for the Tetris domain in order to obtain the best score posible. 
This agent learn with the implementation of the Q-Learning algorithm.
I made this agent as the final project of the subject Artificial Intelligence at College.

All the files are the ones provided by the 2009 rl-competition except for the code that I wrote, which is in the file: 
agents/tetrisAgentJava/src/QTetrisAgent.java
this program also created the file: 
agents/tetrisAgentJava/Qvalues.dat  
where I saved the values of the Q-Matrix obtained during the training.