aiplan4eu/unified-planning

What it solves

It simplifies the process of modeling and solving automated planning problems. Instead of committing to a specific planning approach or language from the start, users can define problems in a planner-independent way and let the library automatically find a compatible solver.

How it works

The library provides a Python API to define the initial state, goals, and possible actions of a system. It can then invoke native solvers or PDDL (Planning Domain Definition Language) planners to find a sequence of actions that achieves the goal. It also supports transformations like grounding, simplification, and the removal of conditional effects, as well as the ability to parse and export problems into PDDL or ANML formats.

Who it’s for

Researchers and developers working in automated planning and AI, who need a flexible way to model complex problems without being tied to a specific solver engine.

Highlights

• Planner-independent problem formulation
• Support for native solvers and PDDL planners
• Automatic engine selection based on problem features
• Ability to dump and parse PDDL and ANML formats
• Built-in problem transformations (e.g., grounding and simplification)