DEC-MAPF — Initial alpha

DEC-MAPF is a Python simulation and experimentation framework for Decentralized Multi-Agent Path Finding. Run studies from Python or the CLI, compare decentralized and centralized methods on shared scenarios, and inspect recorded decisions in the optional visual workspace.

This initial alpha includes:

• Headless experiments with explicit inputs, bounded workers, resumable batches, independent trajectory validation and portable results.
• HeatMap and PathAware negotiation methods, Python CBS and Prioritized Planning, and optional external EECBS/CBSH2-RTC adapters.
• A solver registration interface for method development and shared execution, validation and analysis tools.
• A packaged GUI for scenario configuration, replay, agent and negotiation inspection, paired comparison and export.
• Installation guides, executable study examples, a recorded 80-agent gallery, and software and third-party license notices.

Documentation · Installation · Scientific scope

The release downloads provide a Python wheel with the built GUI, a source distribution with the dependency lockfile, an offline documentation archive and SHA-256 checksums. The package version is 0.1.0a1; the project and repository name is DEC-MAPF. Python imports and CLI commands use mapf.

The wheel can run without Node. GUI dependencies are optional. Native solver binaries are separate tools with their own installation and licensing requirements. PyPI publication is not assumed.

Alpha status describes the software release. Software acceptance checks and the selected valid replay examples do not establish complete reproduction of the articles' statistical results; consult the documented scientific scope when designing comparisons.