planl

Planl is a domain-specific language for expressing planning problems in artificial intelligence. It provides a compact, readable syntax for defining the components of planning problems and can be translated into executable plans by a range of planners. Planl emphasizes declarative descriptions over procedural code, supporting goals, actions, preconditions, effects, and simple temporal constraints. It is intended as an educational and prototyping tool, helping researchers and students compare planning formalisms without heavy boilerplate.

Core concepts in Planl include sorts or types, predicates describing the current state, actions that modify the state, and goals that specify desirable end states. A Planl plan consists of actions arranged to transform the initial state into a goal state. The language supports parameters, preconditions, additive and delete effects, and optional temporal constraints that express when actions may occur.

History and reception: Planl emerged in AI planning research in the late 2010s as a lightweight alternative to more verbose formalisms. It borrows ideas from STRIPS and PDDL while prioritizing readability and ease of parsing for teaching and rapid prototyping. It has enjoyed limited adoption in published work and educational materials, with several open-source implementations that translate Planl descriptions to classical planners.

Example (simplified) Planl description:

action move

parameters: ?r ?from ?to

preconditions: at(?r, ?from)

effects: not at(?r, ?from); at(?r, ?to)

action pick

parameters: ?r ?obj ?loc

preconditions: at(?r, ?loc) and at(?obj, ?loc)

effects: not at(?obj, ?loc); holding(?r, ?obj)

In current practice, Planl is chiefly used as an instructional tool and a stepping stone to standard planning languages. Limitations include restricted expressive power for complex temporal or probabilistic planning and limited standardized tooling across communities.