Exploiting Classical Planning Grounding in Hybrid PDDL+ Planning Engines

Enrico Scala, Mauro Vallati

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

Hybrid PDDL+ models are amongst the most advanced models of systems and the resulting problems are notoriously difficult for planners to cope with due to nonlinear behaviours and immense search spaces. This difficulty is exacerbated by the potentially huge size of the fully ground representations that are used by modern planners in order to effectively explore the search space, which can make some problems impossible to tackle, with the result that in several situations the grounding phase has to be done externally or manually. This not only produces a much less compact problem description, but also complicates debugging and model reuse. To overcome the aforementioned limit, in this paper we investigate two simple grounding techniques for PDDL+ problems. The former method we propose extends the simple mechanism of invariance analysis to limit the groundings of operators upfront. The latter proposes to tackle the grounding process by means of a PDDL+ to Classical Planning abstraction. A preliminary experimental analysis over benchmarks coming from real case study shows that not only the grounding can be sped up, but that also problems that were out of the reach before can now be efficiently solved in an automated manner.
Original languageEnglish
Title of host publicationProceedings - IEEE 32nd International Conference on Tools with Artificial Intelligence, ICTAI 2020
EditorsMiltos Alamaniotis, Shimei Pan
PublisherIEEE Computer Society
Pages85-92
Number of pages8
ISBN (Electronic)9781728192284
ISBN (Print)9781728185361
DOIs
Publication statusPublished - 24 Dec 2020
Event32nd International Conference on Tools with Artificial Intelligence - Online due to COVID-19, Virtual, Baltimore, United States
Duration: 9 Nov 202011 Nov 2020
Conference number: 32
https://ictai2020.org/index.html

Publication series

NameInternational Conference on Tools with Artificial Intelligence, ICTAI
PublisherIEEE
ISSN (Print)1082-3409
ISSN (Electronic)2375-0197

Conference

Conference32nd International Conference on Tools with Artificial Intelligence
Abbreviated titleICTAI 2020
Country/TerritoryUnited States
CityVirtual, Baltimore
Period9/11/2011/11/20
Internet address

Fingerprint

Dive into the research topics of 'Exploiting Classical Planning Grounding in Hybrid PDDL+ Planning Engines'. Together they form a unique fingerprint.

Cite this