Overview

code

The idea of this project is to apply reinforcement learning model-based techniques to a Multi-Agent system. The goal is that the agents try to learn a model of the environment while also learning their own policy.

This “synthetic environment” can be exploited in different ways according to the type of approach selected.

Problem Definition

Consider ‘N’ points on a 2D plane which need to be visited at least once by ‘M’ bots. Provide a path for all of the ‘M’ bots such that the total distance traveled by the bots is minimum.