Skip to main navigation Skip to search Skip to main content

I Always Told My Mom That Order Is Overrated: Unordered HTN Planning is in PSPACE and Models Problems Beyond STRIPS

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

Abstract

Hierarchical Task Network planning is (in)famous for offering great modeling power but making plan existence undecidable. To lift the computational barrier, it is common to enforce orders between all tasks, which allows for an EXPTIME computation. We show that the opposite extreme, using no ordering constraints, reduces complexity further to PSPACE. This unites benefits of STRIPS and HTN planning in a single formalism: It matches the complexity of STRIPS planning while provably giving modelers more expressive power. We observe the same benefit for more, partially new, planning formalisms along the way to the main result. This motivates that these formalisms have similar merit on their own.
Original languageEnglish
Title of host publicationProceedings of the Thirty-Sixth International Conference on Automated Planning and Scheduling (ICAPS 2026)
EditorsAmanda Coles, Wheeler Rumi, Sandhya Saisubramanian
Place of PublicationUSA
PublisherAAAI Press
Pages161-170
Number of pages10
Volume36
ISBN (Electronic)978-1-57735-910-4
ISBN (Print)1-57735-910-0
DOIs
Publication statusPublished - 2026
EventThirty-Sixth International Conference on Automated Planning and Scheduling - Dublin, Ireland
Duration: 27 Jun 20262 Jul 2026

Publication series

NameProceedings of the International Conference on Automated Planning and Scheduling
PublisherAAAI Press
Number1
Volume36
ISSN (Print)2334-0835
ISSN (Electronic)2334-0843

Conference

ConferenceThirty-Sixth International Conference on Automated Planning and Scheduling
Abbreviated titleICAPS 2026
Country/TerritoryIreland
CityDublin
Period27/06/262/07/26

Fingerprint

Dive into the research topics of 'I Always Told My Mom That Order Is Overrated: Unordered HTN Planning is in PSPACE and Models Problems Beyond STRIPS'. Together they form a unique fingerprint.

Cite this