Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Local Guidance for Configuration-Based Multi-Agent Pathfinding
Tomoki Arita, Keisuke Okumura
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add local guidance (LG) to LaCAM | LaCAM · maze-128-128-10, 1,000 agents | flowtime cost: 0.0% -> 38.1% improvement | Improved |
| add global guidance (GG) to LaCAM | LaCAM · maze-128-128-10, 1,000 agents | flowtime cost: 0.0% -> 16.9% improvement | Improved |
| combine local and global guidance (LG+GG) instead of LG alone (ablation) | LaCAM · maze-128-128-10, 1,000 agents | flowtime cost: 38.1% -> 38.4% | Improved |
| use LG instead of GG on warehouse-20-40-10-2-1 | LaCAM · warehouse-20-40-10-2-1 | flowtime | Hurt |
| use LG instead of LNS2 | suboptimal MAPF · MAPF benchmark random scenarios | flowtime and runtime | Improved |
| scale LG to 10,000 agents | LaCAM · warehouse-20-40-10-2-2 | flowtime cost: about 30% cost reduction | Improved |
| cache previous guidance and refine once instead of planning from scratch (ablation) | LG · random-32-32-20 | cost reduction | Improved |
| refine guidance more iterations (ablation) | LG · random-32-32-20 | cost reduction | No clear effect |
| enlarge window size (ablation) | LG · random-32-32-20, warehouse-10-20-10-2-1 | cost and runtime | Mixed |
| overweight collision cost alpha (ablation) | LG · random-32-32-20, warehouse-10-20-10-2-1 | solution quality | Hurt |
| sort agents by collision count in guidance construction (ablation) | LG · empty-32-32, random-32-32-10 | cost reduction | Improved |
| sort agents by collision count (ablation) | LG · warehouse-10-20-10-2-1 | cost reduction | Hurt |
Closest papers by meaning
- Online Guidance Graph Optimization for Lifelong Multi-Agent Path Finding (AAAI 2025 · similarity 0.92)
- Traffic Flow Optimisation for Lifelong Multi-Agent Path Finding (AAAI 2024 · similarity 0.90)
- Learn to Follow: Decentralized Lifelong Multi-Agent Pathfinding via Planning and Learning (AAAI 2024 · similarity 0.90)
- Decentralized Monte Carlo Tree Search for Partially Observable Multi-Agent Pathfinding (AAAI 2024 · similarity 0.89)
- Windowed MAPF with Completeness Guarantees (AAAI 2025 · similarity 0.89)
- Concurrent Planning and Execution in Lifelong Multi-Agent Path Finding with Delay Probabilities (AAAI 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
