Generating Macro-operators by Exploiting Inner Entanglements
Part of book or chapter of book
Kitchin, Diane E.
McCluskey, Thomas Leo
- Publisher: AAAI Press
In Automated Planning, learning and exploiting additional\ud knowledge within a domain model, in order\ud to improve plan generation speed-up and increase\ud the scope of problems solved, has attracted much research.\ud Reformulation techniques such as those based\ud on macro-operators or entanglements are very promising\ud because they are to some extent domain model and\ud planning engine independent. This paper aims to exploit\ud recent work on inner entanglements, relations between\ud pairs of planning operators and predicates encapsulating\ud exclusivity of predicate ‘achievements‘ or ‘requirements’,\ud for generating macro-operators. We provide\ud a theoretical study resulting in a set of conditions\ud when planning operators in an inner entanglement relation\ud can be removed from a domain model and replaced\ud by a macro-operator without compromising solvability\ud of a given (class of) problem(s). The effectiveness of\ud our approach will be experimentally shown on a set\ud of well-known benchmark domains using several highperforming\ud planning engines.
27 references, page 1 of 3
views in local repository
downloads in local repository
The information is available from the following content providers: