Cooperative Game-based Multi-Agent Path Planning with Obstacle Avoidance
Yaning Guo, Quan Pan, Qi Yuan Sun, Kai Zhao, Dong Wang, Min Feng · 2019
This paper investigates multi-agent cooperative path planning with obstacle avoidance based on game theory and multi-agent reinforcement learning algorithm. It aims to extend the traditional single agent Q-learning algorithm to multi-agent systems by using the cooperative game framework. This framework takes into account the selection of joint actions at joint states for multi-agent cooperative path planning with obstacle avoidance. First, a cooperative game model is presented for agents to achieve cooperative path planning with obstacle avoidance in complicated environment. Second, a multi-agent Q-learning algorithm in continuous state space is proposed for solving Nash equilibrium, where the local minimum problem is well resolved. Finally, a numerical example is conducted to verify the effectiveness of the proposed approach.