Extending the use of plateau-escaping macro-actions in planning

Amanda Smith

Research output: Contribution to conferencePaper

Abstract

Many fully automated planning systems use a single, domain independent heuristic to guide search and no other problem specific guidance. While these systems exhibit excellent performance, they are often out-performed by systems which are either given extra human-encoded search information, or spend time learning additional search control information offline. The benefit of systems which do not require human intervention is that they are much closer to the ideal of autonomy. This document discusses a system which learns additional control knowledge, in the form of macro-actions, during planning, without the additional time required for an online learning step. The results of various techniques for managing the collection of macro-actions generated are also discussed. Finally, an explanation of the extension of the techniques to other planning systems is presented.

Conference

ConferenceInternational Conference on Automated Planning and Scheduling (ICAPS)
CityCumbria, UK
Period6/06/0610/06/06

Fingerprint

Macros
Planning

Keywords

  • planning systems
  • plateaus
  • search engines
  • control systems
  • automated planning systems

Cite this

Smith, A. (2006). Extending the use of plateau-escaping macro-actions in planning. Paper presented at International Conference on Automated Planning and Scheduling (ICAPS), Cumbria, UK, .
Smith, Amanda. / Extending the use of plateau-escaping macro-actions in planning. Paper presented at International Conference on Automated Planning and Scheduling (ICAPS), Cumbria, UK, .
@conference{df68603f0d41429f847fde341fc7b355,
title = "Extending the use of plateau-escaping macro-actions in planning",
abstract = "Many fully automated planning systems use a single, domain independent heuristic to guide search and no other problem specific guidance. While these systems exhibit excellent performance, they are often out-performed by systems which are either given extra human-encoded search information, or spend time learning additional search control information offline. The benefit of systems which do not require human intervention is that they are much closer to the ideal of autonomy. This document discusses a system which learns additional control knowledge, in the form of macro-actions, during planning, without the additional time required for an online learning step. The results of various techniques for managing the collection of macro-actions generated are also discussed. Finally, an explanation of the extension of the techniques to other planning systems is presented.",
keywords = "planning systems, plateaus, search engines, control systems, automated planning systems",
author = "Amanda Smith",
year = "2006",
month = "6",
day = "6",
language = "English",
note = "International Conference on Automated Planning and Scheduling (ICAPS) ; Conference date: 06-06-2006 Through 10-06-2006",

}

Smith, A 2006, 'Extending the use of plateau-escaping macro-actions in planning' Paper presented at International Conference on Automated Planning and Scheduling (ICAPS), Cumbria, UK, 6/06/06 - 10/06/06, .

Extending the use of plateau-escaping macro-actions in planning. / Smith, Amanda.

2006. Paper presented at International Conference on Automated Planning and Scheduling (ICAPS), Cumbria, UK, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - Extending the use of plateau-escaping macro-actions in planning

AU - Smith, Amanda

PY - 2006/6/6

Y1 - 2006/6/6

N2 - Many fully automated planning systems use a single, domain independent heuristic to guide search and no other problem specific guidance. While these systems exhibit excellent performance, they are often out-performed by systems which are either given extra human-encoded search information, or spend time learning additional search control information offline. The benefit of systems which do not require human intervention is that they are much closer to the ideal of autonomy. This document discusses a system which learns additional control knowledge, in the form of macro-actions, during planning, without the additional time required for an online learning step. The results of various techniques for managing the collection of macro-actions generated are also discussed. Finally, an explanation of the extension of the techniques to other planning systems is presented.

AB - Many fully automated planning systems use a single, domain independent heuristic to guide search and no other problem specific guidance. While these systems exhibit excellent performance, they are often out-performed by systems which are either given extra human-encoded search information, or spend time learning additional search control information offline. The benefit of systems which do not require human intervention is that they are much closer to the ideal of autonomy. This document discusses a system which learns additional control knowledge, in the form of macro-actions, during planning, without the additional time required for an online learning step. The results of various techniques for managing the collection of macro-actions generated are also discussed. Finally, an explanation of the extension of the techniques to other planning systems is presented.

KW - planning systems

KW - plateaus

KW - search engines

KW - control systems

KW - automated planning systems

UR - http://www.cis.strath.ac.uk/research/publications/papers/strath_cis_publication_1546.pdf

M3 - Paper

ER -

Smith A. Extending the use of plateau-escaping macro-actions in planning. 2006. Paper presented at International Conference on Automated Planning and Scheduling (ICAPS), Cumbria, UK, .