author: Ronald Meijer
title: PDDL Planning Problems and GROOVE Graph Transformations: Combining two Worlds with a Translator
keywords:
topics: Graphs
committee: Arend Rensink
end: June 2012

Abstract

This paper describes how the AI planning language PDDL was combined with a graph-based model checking tool called GROOVE. Graph-based models are very intuitive to work with, yet powerful enough to deal with a great variety of problems. The visual representation provides a quick insight in the workings of a model and the available tools allows the user to explore the system with great ease. The built-in state space generator of GROOVE comes in handy during the analysis of a problem and its solution.

Traditional PDDL planners on the other hand are specialized in finding a solution without exploring the complete state space. This approach can be necessary for problems which state space is just too big to fully explore. Both approaches have shown to be useful, and the translator we built combines the forces of both: it makes for an easy interpretation of traditional planning problems formulated in PDDL, and exploration process for graph transformation systems with a big state space can be accelerated with an export to PDDL.

Additional Resources

  1. The paper