VI51 Project - UTBM student project
Artificial Intelligence for lemmings based on Q-Learning, multi-agent approach.
Authors : ENNAJI Floriane, GOMEZ Lucille, THIBAUD Romain, WALTZ Antonin
##Resources needed
- Eclipse SARL (v0.3.1): http://www.sarl.io/download/index.html (last visit 12/06/2016)
- Janus platform : http://www.janusproject.io/ (last visit (12/06/2016)
##Before running
- Import the Janus library with the external jar downloaded (http://www.janusproject.io/) and remove the Sarl default library
##Execution There're two ways to execute the application :
####Learning mode It will create a QTable that will be used in the auto-playing mode.
To enable it, go to Game.java
and ensure that main
function looks like following :
/*
* Comment following if you want to play
*/
launchLearning(worlds, 3);
/*
* Uncomment following if you want to play
*/
/*Random rand = new Random();
int worldNb = rand.nextInt(worlds.length);
File file = new File(worlds[worldNb]);
BufferedImage image = ImageIO.read(file);
env = new Environment(image);
QTable qt = getQTableFromFile();
env.setQTable(qt);
new Launcher(env);*/
####Auto-playing mode
It consists in letting Lemming use the QTable to live in the world. To enable it, ensure that main
function looks like following :
/*
* Comment following if you want to play
*/
/*launchLearning(worlds, 3);*/
/*
* Uncomment following if you want to play
*/
Random rand = new Random();
int worldNb = rand.nextInt(worlds.length);
File file = new File(worlds[worldNb]);
BufferedImage image = ImageIO.read(file);
env = new Environment(image);
QTable qt = getQTableFromFile();
env.setQTable(qt);
new Launcher(env);