Planning for multiple goals with limited interactions.

Dana S. Nau*, Qiang Yang, James Hendler

*Corresponding author for this work

Research output: Chapter in Book/Conference Proceeding/ReportConference Paper published in a bookpeer-review

1 Citation (Scopus)

Abstract

A description is presented of the problem of finding an optimal global plan for a multigoal planning problem by combining plans that solve the individual goals. This problem arises in process planning, when one tries to combine the process plans for individual machinable features into a global process plan for the entire workpiece. It also arises in the planning of robot actions, when one tries to combine the plans for the handling of individual objects into a global plan. The problem can show to be NP-hard in general. However, if only one plan is available for each goal, then the problem can be solved in time O(n2) by imposing restrictions on the kinds of intergoal interactions involved. If more than one plan is available for each goal, the problem is still NP-hard even if these restrictions are satisfied, but in this case a heuristic search algorithm has been developed which performs well.

Original languageEnglish
Title of host publicationProc Fifth Conf Artif Intell
Editors Anon
PublisherPubl by IEEE
Pages263-270
Number of pages8
ISBN (Print)0818619023
Publication statusPublished - 1988
Externally publishedYes
EventProceedings - Fifth Conference on Artificial Intelligence Applications - Miami, FL, USA
Duration: 6 Mar 198910 Mar 1989

Publication series

NameProc Fifth Conf Artif Intell

Conference

ConferenceProceedings - Fifth Conference on Artificial Intelligence Applications
CityMiami, FL, USA
Period6/03/8910/03/89

Fingerprint

Dive into the research topics of 'Planning for multiple goals with limited interactions.'. Together they form a unique fingerprint.

Cite this