Multi-Agent Collaborative Reinforcement Learning

Multi-agent systems are gaining popularity. In this thesis Multi-agent box pushing task is considered. This problem provides challenges in collaboration while keeping state-action space complexity to minimum. The task is to push a box from source to goal. This task is executed by multi-agents always adjacent to box. To efficiently solve problems in a multi-agent framework machine learning is required. In this thesis reinforcement learning is used to solve multi-agent box pushing task.Keywords: Reinforcement learning; Q-learning; Game Theory; Multi-Agent System.Best resource for beginners in the Reinforcement Learning field. Read more