/Reinforcement-Learning-simulation-Java

QLearning implementation reinforced by norms in a multi-agent simulation

Primary LanguageJava

NormRLJava

Project

This project purpose is to simulate how agents can learn individually to build patterns and collective behaviors to survive and achieve their goals. These behaviors can emerge from the implementation of Q-Learning, improved by the building of a norms. The program represents a roundabout simulation, where agents, that are cars, have to find their way in the most efficient way possible while avoiding collision with other agents.

Execution

You can download the file Simulation.jar To execute it, go into the file folder, then type in the console the following line: java -jar Simulation.jar

You will then be able to chose the parameters for the simulation.

Javadoc

The full english documentation of the program is furnished in the /doc folder