The paper introduces PDDL2.1, an extension of the Planning Domain Description Language (PDDL) designed to handle temporal and resource-intensive planning domains. The authors, Maria Fox and Derek Long, highlight the growing importance of planning in realistic applications involving time and resources, such as logistics, plant control, and manufacturing. The International Planning Competitions have driven the development of more expressive modeling languages, and PDDL2.1 was developed to meet the challenges posed by these competitions.
PDDL2.1 retains backward compatibility with the original PDDL, ensuring that existing domains and plans remain valid. It introduces numeric expressions and durative actions to enhance the language's expressive power. Numeric expressions allow for the modeling of resources and metrics, enabling planners to evaluate plans based on specific criteria. Durative actions, both discretized and continuous, enable the modeling of temporal properties and concurrent activities, though with certain constraints to ensure plan validity.
The paper discusses the syntax, semantics, and validation of concurrent plans, emphasizing the importance of distinguishing between conditions and effects at the start and end points of durative actions. It also addresses the challenges of concurrency and resource management, particularly in the context of numeric changes and discrete updates.
Overall, PDDL2.1 aims to provide a robust and flexible language for expressing complex planning domains, while maintaining compatibility with existing planning systems and research communities.The paper introduces PDDL2.1, an extension of the Planning Domain Description Language (PDDL) designed to handle temporal and resource-intensive planning domains. The authors, Maria Fox and Derek Long, highlight the growing importance of planning in realistic applications involving time and resources, such as logistics, plant control, and manufacturing. The International Planning Competitions have driven the development of more expressive modeling languages, and PDDL2.1 was developed to meet the challenges posed by these competitions.
PDDL2.1 retains backward compatibility with the original PDDL, ensuring that existing domains and plans remain valid. It introduces numeric expressions and durative actions to enhance the language's expressive power. Numeric expressions allow for the modeling of resources and metrics, enabling planners to evaluate plans based on specific criteria. Durative actions, both discretized and continuous, enable the modeling of temporal properties and concurrent activities, though with certain constraints to ensure plan validity.
The paper discusses the syntax, semantics, and validation of concurrent plans, emphasizing the importance of distinguishing between conditions and effects at the start and end points of durative actions. It also addresses the challenges of concurrency and resource management, particularly in the context of numeric changes and discrete updates.
Overall, PDDL2.1 aims to provide a robust and flexible language for expressing complex planning domains, while maintaining compatibility with existing planning systems and research communities.