mats-lp

[AAAI-2024] MATS-LP addresses the challenging problem of decentralized lifelong multi-agent pathfinding. The proposed approach utilizes a combination of Monte Carlo Tree Search and reinforcement learning for resolving conflicts.