Skip to content
@openplan-labs

OpenPlan Labs

Open tools for automated planning — CUDA-accelerated and classical planners, multi-agent pathfinding, and PDDL tooling.
README.md
OpenPlan Labs

OpenPlan Labs

Open tools for automated planning — CUDA-accelerated and classical planners, multi-agent pathfinding, and PDDL tooling.

Everything here is an implementation of published work, written to be read as well as run. Each README says what the tool does and what it does not do; if one doesn't, that is a bug worth filing.

Multi-agent pathfinding

RepositoryWhat it is
pymapfPython library for multi-agent planning and pathfinding, centralized and decentralized. On PyPI.
cuda-planningThe same algorithms batched onto a GPU, with a NumPy reference backend. Installs as cuplan — the repository and the distribution are deliberately named differently.

PDDL & classical planning

RepositoryWhat it is
PythonPDDLDependency-free, pure-Python PDDL planning framework: parser, grounder, planners, heuristics, and a benchmarking harness. Installs as jupyddl.
pddl-examplesCurated PDDL domains and problems. Planners not included. Contributions of new instances are welcome.
awesome-pddlCurated list of PDDL resources — planners, editors, tutorials, papers.
docker-pythonpddlPythonPDDL in a container, for running without a local Python setup.

Measurement & figures

RepositoryWhat it is
openplan-benchOne harness, one problem set, one results schema for every planner here. Timeouts and errors are recorded as rows, not dropped. Leaderboard.
planvizGrid maps, search animations, plan timelines and benchmark charts, light and dark from the same call. Planners not included — it draws results, it does not produce them. On PyPI.

Meta

RepositoryWhat it is
brandingThe Frontier brand system — palette, mark, type, figure conventions. If a repo's colours drift from it, it wins.
.githubOrg-wide defaults: contributing guide, code of conduct, security policy, issue templates.

Get involved

Bug reports and PDDL instances are the easiest ways in; both have templates. The contributing guide covers the PR flow, and each project's own README covers its specifics. Security reports go privately to the address in the security policy.

Pinned Loading

  1. pymapfpymapfPublic

    A Python library for multi-agent planning and pathfinding: CBS, PIBT, LaCAM, LNS, space-time A*, plus decentralized NMPC, velocity obstacles and swarm behaviors

    Python 77 12

  2. PythonPDDLPythonPDDLPublic

    A dependency-free, pure-Python PDDL planning framework: parser, grounder, classical-to-SOTA planners, heuristics and a benchmarking harness

    Python 36 3

  3. pddl-examplespddl-examplesPublic

    Curated and auto-synced collection of PDDL domains and problems

    PDDL 3

  4. cuda-planningcuda-planningPublic

    CUDA-accelerated multi-agent path finding and planning: batched BFS, space-time search, prioritized planning, PIBT, velocity obstacles, flocking — with a NumPy reference backend

    Python

  5. openplan-benchopenplan-benchPublic

    Reproducible benchmarking harness and living leaderboard for the OpenPlan Labs planners — classical PDDL and MAPF, with timeouts and errors recorded as results

    Python

  6. planvizplanvizPublic

    Frontier-styled matplotlib figures for planning and search: grid maps, search animations, plan timelines, and benchmark charts

    Python

Repositories

Showing 10 of 11 repositories
View all repositories

Top languages

Loading…

Most used topics

Loading…