Enhancing Lifelong Multi-Agent Path-finding by Using Artificial Potential Fields

2025-05-28

AI Summary

This work addresses the Lifelong Multi-Agent Path Finding (LMAPF) problem under dynamic, continuous-task settings—where agents operate indefinitely while goals are updated online, requiring conflict-free, high-throughput coordinated navigation. We propose the first systematic integration of Artificial Potential Fields (APF) into the LMAPF framework, overcoming APF’s known failure in classical discrete MAPF by synergistically combining it with priority-based planning, MAPF-LNS2, and PIBT to form a hybrid, dynamically adaptive planner. Our core innovation lies in leveraging APF for continuous collision avoidance and goal-directed guidance, thereby significantly improving path re-planning efficiency and robustness. Experiments on standard LMAPF benchmarks demonstrate up to a 7× improvement in long-term system throughput, validating APF’s intrinsic benefits and practical utility in lifelong multi-agent navigation.

Technology Category

• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

We explore the use of Artificial Potential Fields (APFs) to solve Multi-Agent Path Finding (MAPF) and Lifelong MAPF (LMAPF) problems. In MAPF, a team of agents must move to their goal locations without collisions, whereas in LMAPF, new goals are generated upon arrival. We propose methods for incorporating APFs in a range of MAPF algorithms, including Prioritized Planning, MAPF-LNS2, and Priority Inheritance with Backtracking (PIBT). Experimental results show that using APF is not beneficial for MAPF but yields up to a 7-fold increase in overall system throughput for LMAPF.

Problem

• Enhancing Lifelong MAPF using Artificial Potential Fields
• Solving collision-free pathfinding for multi-agent systems
• Improving system throughput in dynamic goal scenarios

Innovation

• Artificial Potential Fields for MAPF solutions
• APF integration in Prioritized Planning algorithms
• APF boosts LMAPF throughput sevenfold