Reproduce “Multi-Agent Path Finding via Finite-Horizon Hierarchical Factorization” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiarui Li, Alessandro Zanardi, Gioele Zardini
An online, parallelizable Multi-Agent Path Finding (MAPF) algorithm is introduced, utilizing finite-horizon hierarchical factorization to coordinate large robot fleets. This method substantially reduces planning delays and achieves superior solution quality by enabling immediate execution and efficiently handling conflicts through dynamic grouping and parallel replanning.
Key insights
- Decomposing the global planning problem into smaller, time-limited subproblems processed in a receding-horizon manner allows for immediate execution and improved responsiveness.
- Extensive parallelization at both individual and group levels, combined with dynamic grouping based on conflict reachability, enables efficient scaling to a large number of agents.
- A finite planning horizon effectively manages computational load, while the optimal horizon length varies depending on map structure and agent density, impacting the balance between planning efficiency and solution quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.