Multi-Agent Simulation

<h1>Multi-agent simulation : cellular automaton and boids (Java)</h1>

Contributors: Mathieu Tillet, Aymeric Lhullier

This project was part of the Object-Oriented Programming course. The aim was to develop in Java an application for graphically simulating multi-agent systems. Simulated systems included three systems of the cellular automaton type:

• Conway's game of life
• Immigration game
• Schelling's segregation model

Another simulated system was a self-organized swarm movement system: the Boids model.

Click the image to see a small demo of the result:

I was responsible for:

• Global class structuring of the project
• Event manager : update rate, boid death's...
• Schelling's segregation model implementation
• Wolf boid AI

Cellular automatons

Each cellular automaton implemented follows specific update rules but share the remaining implementation thanks to inheritance.

Boids

The Boids model is a flocking simulation model. Each agent is subject to at least two constraints :

• Neighboor effect : depending on their proximity, neighboors of the same specie attract or repel (if too close) each other.
• Screen encasing : borders of the screen repel the boids so as they stay visible during the simulation.

Specific species of boids were created, with customized behaviours :

• Wolves (in red) : wolves tend to hunt together, are pretty fast, are attracted to sheeps and afraid of herdsmen.
• Sheeps (in blue) : sheeps like to stay in big herds and regroup behind herdsmen. They are afraid of wolves but slower than them.
• Herdsmen (in green) : herdsmen try to protect sheeps from wolves. They can kill nearby wolves but are much slower than the other boids.