Learning macro-actions for arbitrary planners and domains

M.A. Hakim Newton, John Levine, Maria Fox, Derek Long, Mark Boddy (Editor), Maria Fox (Editor), Sylvie Thiebaux (Editor)

Research output: Chapter in Book/Report/Conference proceedingChapter

47 Citations (Scopus)

Abstract

Many complex domains and even larger problems in simple domains remain challenging in spite of the recent progress in planning. Besides developing and improving planning technologies, re-engineering a domain by utilising acquired knowledge opens up a potential avenue for further research. Moreover, macro-actions, when added to the domain as additional actions, provide a promising means by which to convey such knowledge. A macro-action, or macro in short, is a group of actions selected for application as a single choice. Most existing work on macros exploits properties explicitly specific to the planners or the domains. However, such properties are not likely to be common with arbitrary planners or domains. Therefore, a macro learning method that does not exploit any structural knowledge about planners or domains explicitly is of immense interest. This paper presents an offline macro learning method that works with arbitrarily chosen planners and domains. Given a planner, a domain, and a number of example problems, the learning method generates macros from plans of some of the given problems under the guidance of a genetic algorithm. It represents macros like regular actions, evaluates them individually by solving the remaining given problems, and suggests individual macros that are to be added to the domain permanently. Genetic algorithms are automatic learning methods that can capture inherent features of a system using no explicit knowledge about it. Our method thus does not strive to discover or utilise any structural properties specific to a planner or a domain.
LanguageEnglish
Title of host publicationProceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007)
Place of PublicationCalifornia, USA
Pages256-263
Number of pages7
Publication statusPublished - 2007

Fingerprint

Macros
Genetic algorithms
Planning
Structural properties

Keywords

  • macro-actions
  • domain planning
  • domain engineering
  • offline macro learning method

Cite this

Newton, M. A. H., Levine, J., Fox, M., Long, D., Boddy, M. (Ed.), Fox, M. (Ed.), & Thiebaux, S. (Ed.) (2007). Learning macro-actions for arbitrary planners and domains. In Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007) (pp. 256-263). California, USA.
Newton, M.A. Hakim ; Levine, John ; Fox, Maria ; Long, Derek ; Boddy, Mark (Editor) ; Fox, Maria (Editor) ; Thiebaux, Sylvie (Editor). / Learning macro-actions for arbitrary planners and domains. Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007). California, USA, 2007. pp. 256-263
@inbook{ad9aaf2c27ad437b9f190f6b524b5177,
title = "Learning macro-actions for arbitrary planners and domains",
abstract = "Many complex domains and even larger problems in simple domains remain challenging in spite of the recent progress in planning. Besides developing and improving planning technologies, re-engineering a domain by utilising acquired knowledge opens up a potential avenue for further research. Moreover, macro-actions, when added to the domain as additional actions, provide a promising means by which to convey such knowledge. A macro-action, or macro in short, is a group of actions selected for application as a single choice. Most existing work on macros exploits properties explicitly specific to the planners or the domains. However, such properties are not likely to be common with arbitrary planners or domains. Therefore, a macro learning method that does not exploit any structural knowledge about planners or domains explicitly is of immense interest. This paper presents an offline macro learning method that works with arbitrarily chosen planners and domains. Given a planner, a domain, and a number of example problems, the learning method generates macros from plans of some of the given problems under the guidance of a genetic algorithm. It represents macros like regular actions, evaluates them individually by solving the remaining given problems, and suggests individual macros that are to be added to the domain permanently. Genetic algorithms are automatic learning methods that can capture inherent features of a system using no explicit knowledge about it. Our method thus does not strive to discover or utilise any structural properties specific to a planner or a domain.",
keywords = "macro-actions, domain planning, domain engineering, offline macro learning method",
author = "Newton, {M.A. Hakim} and John Levine and Maria Fox and Derek Long and Mark Boddy and Maria Fox and Sylvie Thiebaux",
year = "2007",
language = "English",
isbn = "1577353447",
pages = "256--263",
booktitle = "Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007)",

}

Newton, MAH, Levine, J, Fox, M, Long, D, Boddy, M (ed.), Fox, M (ed.) & Thiebaux, S (ed.) 2007, Learning macro-actions for arbitrary planners and domains. in Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007). California, USA, pp. 256-263.

Learning macro-actions for arbitrary planners and domains. / Newton, M.A. Hakim; Levine, John; Fox, Maria; Long, Derek; Boddy, Mark (Editor); Fox, Maria (Editor); Thiebaux, Sylvie (Editor).

Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007). California, USA, 2007. p. 256-263.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - Learning macro-actions for arbitrary planners and domains

AU - Newton, M.A. Hakim

AU - Levine, John

AU - Fox, Maria

AU - Long, Derek

A2 - Boddy, Mark

A2 - Fox, Maria

A2 - Thiebaux, Sylvie

PY - 2007

Y1 - 2007

N2 - Many complex domains and even larger problems in simple domains remain challenging in spite of the recent progress in planning. Besides developing and improving planning technologies, re-engineering a domain by utilising acquired knowledge opens up a potential avenue for further research. Moreover, macro-actions, when added to the domain as additional actions, provide a promising means by which to convey such knowledge. A macro-action, or macro in short, is a group of actions selected for application as a single choice. Most existing work on macros exploits properties explicitly specific to the planners or the domains. However, such properties are not likely to be common with arbitrary planners or domains. Therefore, a macro learning method that does not exploit any structural knowledge about planners or domains explicitly is of immense interest. This paper presents an offline macro learning method that works with arbitrarily chosen planners and domains. Given a planner, a domain, and a number of example problems, the learning method generates macros from plans of some of the given problems under the guidance of a genetic algorithm. It represents macros like regular actions, evaluates them individually by solving the remaining given problems, and suggests individual macros that are to be added to the domain permanently. Genetic algorithms are automatic learning methods that can capture inherent features of a system using no explicit knowledge about it. Our method thus does not strive to discover or utilise any structural properties specific to a planner or a domain.

AB - Many complex domains and even larger problems in simple domains remain challenging in spite of the recent progress in planning. Besides developing and improving planning technologies, re-engineering a domain by utilising acquired knowledge opens up a potential avenue for further research. Moreover, macro-actions, when added to the domain as additional actions, provide a promising means by which to convey such knowledge. A macro-action, or macro in short, is a group of actions selected for application as a single choice. Most existing work on macros exploits properties explicitly specific to the planners or the domains. However, such properties are not likely to be common with arbitrary planners or domains. Therefore, a macro learning method that does not exploit any structural knowledge about planners or domains explicitly is of immense interest. This paper presents an offline macro learning method that works with arbitrarily chosen planners and domains. Given a planner, a domain, and a number of example problems, the learning method generates macros from plans of some of the given problems under the guidance of a genetic algorithm. It represents macros like regular actions, evaluates them individually by solving the remaining given problems, and suggests individual macros that are to be added to the domain permanently. Genetic algorithms are automatic learning methods that can capture inherent features of a system using no explicit knowledge about it. Our method thus does not strive to discover or utilise any structural properties specific to a planner or a domain.

KW - macro-actions

KW - domain planning

KW - domain engineering

KW - offline macro learning method

UR - http://www.aiconferences.org/ICAPS/2007/icaps07.html

M3 - Chapter

SN - 1577353447

SP - 256

EP - 263

BT - Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007)

CY - California, USA

ER -

Newton MAH, Levine J, Fox M, Long D, Boddy M, (ed.), Fox M, (ed.) et al. Learning macro-actions for arbitrary planners and domains. In Proceedings of the Seventeenth International Conference on Automated Planning and Scheduling (ICAPS 2007). California, USA. 2007. p. 256-263