Lifelong LaCAM with Local Guidance for Lifelong MAPF

2026-05-16

AI Summary

This work addresses the challenges of congestion and reduced throughput in lifelong multi-agent pathfinding (LMAPF) caused by continuously arriving tasks. To tackle this, the authors propose LLLG, a locally guided extension of LaCAM that introduces, for the first time, a local spatiotemporal guidance mechanism into lifelong MAPF. LLLG integrates a receding-horizon planning framework with a warm-start strategy, leveraging prior solutions to efficiently guide current planning. The method achieves significantly higher long-term task throughput while maintaining high scalability and strong real-time performance even in dense scenarios. Experimental results demonstrate that LLLG substantially outperforms existing approaches in task completion rates, advancing the state of the art in real-time lifelong MAPF.

Abstract

Local guidance has recently proven to be a powerful driver of empirical performance in real-time, suboptimal multi-agent pathfinding (MAPF), improving the scalable configuration-based solver LaCAM. By injecting informative spatiotemporal cues around each agent, local guidance mitigates congestion, reduces waiting, and remains scalable enough even with tight time budgets, yielding state-of-the-art performance for one-shot MAPF. This study asks whether the same benefits can be lifted to the lifelong setting (LMAPF), where tasks arrive continuously and improvements in per-step plans can increase task completion throughput over long horizons. We propose LLLG, a Lifelong version of LaCAM enhanced with Local Guidance, which employs a receding-horizon windowed planning framework and warm-starts guidance from the previous solution at each timestep. Our method scales effectively, maintains high throughput even in compact, dense environments, and surpasses existing planners, thereby pushing the frontier of real-time, lifelong MAPF.

Problem

• Lifelong MAPF
• Local Guidance
• Multi-Agent Pathfinding
• Throughput Optimization
• Real-Time Planning

Innovation

• Lifelong MAPF
• Local Guidance
• LaCAM
• Receding-horizon Planning
• Multi-agent Pathfinding