More Succinct Grounding of HTN Planning Problems -- Preliminary Results
Gregor Behnke, Daniel Höller, Pascal Bercher, Susanne Biundo · ANU Open Research (Australian National University) · 2019
Planning systems usually operate on grounded representations of the planning problems during search. Further, planners that use translations into other combinatorial problems also often perform their translations based on a grounded model. Planning models, however, are commonly defined in a lifted formalism. As such, one of the first preprocessing steps a planner performs is to generate a grounded representation. In this paper we present a new approach for grounding HTN planning problems that produces smaller groundings than the previously published method. We expect this decrease in size to lead to more efficient planners.