Abstract
AI Planning is inherently hard and hence it is desirable to derive as much information as we can from the structure of the planning problem and let this information be exploited by a planner. Many recent planners use the finite-domain state-variable representation of the problem instead of the classical propositional representation.
However, most planning problems are still specified in the propositional representation due to the widespread modelling language planning domain definition language and it is hard to generate an efficient state-variable representation from the propositional model. In this article, we investigate various methods for automated generation of efficient state-variable representations from the propositional representation and we propose a novel approach - constructed as a combination of existing techniques - that utilises the structural information from the goal and the initial state.
We perform an exhaustive experimental evaluation of methods, planning systems and problems, using the International Planning Competition as the main source of data. We show that for many planning problems the novel approach provides an improved efficiency.