Analysing AI Planning Problems in Linear Logic – A Partial Deduction Approach

This article presents a framework for analysing AI planning problem specifications. We consider AI planning as linear logic (LL) theorem proving. Then the usage of partial deduction is proposed as a foundation of an analysis technique for AI planning problems, which are described in LL. By applying this technique we are able to investigate for instance why there is no solution for a particular planning problem. We consider here !-Horn fragment of LL, which is expressive enough for representing STRIPS-like planning problems. Anyway, by taking advantage of full LL, more expressive planning problems can be described, Therefore, the framework proposed here could be seen as a step towards analysing both, STRIPS-like and more complex planning problems.