Derivation replay for partial-order planning

Laurie H. Ihrig, Subbarao Kambhampati

Research output: Chapter in Book/Report/Conference proceedingConference contribution

11 Citations (Scopus)

Abstract

Derivation replay was first proposed by Carbonell as a method of transferring guidance from a previous problem-solving episode to a new one. Subsequent implementations have used state-space planning as the underlying methodology. This paper is motivated by the acknowledged superiority of partial-order (PO) planners in plan generation, and is an attempt to bring derivation replay into the realm of partial-order planning. Here we develop DerSNLP, a framework for doing replay in SNLP, a partial-order plan-space planner, and analyze its relative effectiveness. We will argue that the decoupling of planning (derivation) order and the execution order of plan steps, provided by partial-order planners, enables DerSNLP to exploit the guidance of previous cases in a more efficient and straightforward fashion. We validate our hypothesis through empirical comparisons between DerSNLP and two replay systems based on state-space planners.

Original languageEnglish (US)
Title of host publicationProceedings of the National Conference on Artificial Intelligence
Place of PublicationMenlo Park, CA, United States
PublisherAAAI
Pages992-997
Number of pages6
Volume2
StatePublished - 1994
EventProceedings of the 12th National Conference on Artificial Intelligence. Part 1 (of 2) - Seattle, WA, USA
Duration: Jul 31 1994Aug 4 1994

Other

OtherProceedings of the 12th National Conference on Artificial Intelligence. Part 1 (of 2)
CitySeattle, WA, USA
Period7/31/948/4/94

Fingerprint

Planning

ASJC Scopus subject areas

  • Software

Cite this

Ihrig, L. H., & Kambhampati, S. (1994). Derivation replay for partial-order planning. In Proceedings of the National Conference on Artificial Intelligence (Vol. 2, pp. 992-997). Menlo Park, CA, United States: AAAI.

Derivation replay for partial-order planning. / Ihrig, Laurie H.; Kambhampati, Subbarao.

Proceedings of the National Conference on Artificial Intelligence. Vol. 2 Menlo Park, CA, United States : AAAI, 1994. p. 992-997.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ihrig, LH & Kambhampati, S 1994, Derivation replay for partial-order planning. in Proceedings of the National Conference on Artificial Intelligence. vol. 2, AAAI, Menlo Park, CA, United States, pp. 992-997, Proceedings of the 12th National Conference on Artificial Intelligence. Part 1 (of 2), Seattle, WA, USA, 7/31/94.
Ihrig LH, Kambhampati S. Derivation replay for partial-order planning. In Proceedings of the National Conference on Artificial Intelligence. Vol. 2. Menlo Park, CA, United States: AAAI. 1994. p. 992-997
Ihrig, Laurie H. ; Kambhampati, Subbarao. / Derivation replay for partial-order planning. Proceedings of the National Conference on Artificial Intelligence. Vol. 2 Menlo Park, CA, United States : AAAI, 1994. pp. 992-997
@inproceedings{219b0dc21538419f968c9498985f5a15,
title = "Derivation replay for partial-order planning",
abstract = "Derivation replay was first proposed by Carbonell as a method of transferring guidance from a previous problem-solving episode to a new one. Subsequent implementations have used state-space planning as the underlying methodology. This paper is motivated by the acknowledged superiority of partial-order (PO) planners in plan generation, and is an attempt to bring derivation replay into the realm of partial-order planning. Here we develop DerSNLP, a framework for doing replay in SNLP, a partial-order plan-space planner, and analyze its relative effectiveness. We will argue that the decoupling of planning (derivation) order and the execution order of plan steps, provided by partial-order planners, enables DerSNLP to exploit the guidance of previous cases in a more efficient and straightforward fashion. We validate our hypothesis through empirical comparisons between DerSNLP and two replay systems based on state-space planners.",
author = "Ihrig, {Laurie H.} and Subbarao Kambhampati",
year = "1994",
language = "English (US)",
volume = "2",
pages = "992--997",
booktitle = "Proceedings of the National Conference on Artificial Intelligence",
publisher = "AAAI",

}

TY - GEN

T1 - Derivation replay for partial-order planning

AU - Ihrig, Laurie H.

AU - Kambhampati, Subbarao

PY - 1994

Y1 - 1994

N2 - Derivation replay was first proposed by Carbonell as a method of transferring guidance from a previous problem-solving episode to a new one. Subsequent implementations have used state-space planning as the underlying methodology. This paper is motivated by the acknowledged superiority of partial-order (PO) planners in plan generation, and is an attempt to bring derivation replay into the realm of partial-order planning. Here we develop DerSNLP, a framework for doing replay in SNLP, a partial-order plan-space planner, and analyze its relative effectiveness. We will argue that the decoupling of planning (derivation) order and the execution order of plan steps, provided by partial-order planners, enables DerSNLP to exploit the guidance of previous cases in a more efficient and straightforward fashion. We validate our hypothesis through empirical comparisons between DerSNLP and two replay systems based on state-space planners.

AB - Derivation replay was first proposed by Carbonell as a method of transferring guidance from a previous problem-solving episode to a new one. Subsequent implementations have used state-space planning as the underlying methodology. This paper is motivated by the acknowledged superiority of partial-order (PO) planners in plan generation, and is an attempt to bring derivation replay into the realm of partial-order planning. Here we develop DerSNLP, a framework for doing replay in SNLP, a partial-order plan-space planner, and analyze its relative effectiveness. We will argue that the decoupling of planning (derivation) order and the execution order of plan steps, provided by partial-order planners, enables DerSNLP to exploit the guidance of previous cases in a more efficient and straightforward fashion. We validate our hypothesis through empirical comparisons between DerSNLP and two replay systems based on state-space planners.

UR - http://www.scopus.com/inward/record.url?scp=0028585768&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0028585768&partnerID=8YFLogxK

M3 - Conference contribution

VL - 2

SP - 992

EP - 997

BT - Proceedings of the National Conference on Artificial Intelligence

PB - AAAI

CY - Menlo Park, CA, United States

ER -