A heuristic approach to planning with incomplete STRIPS action models

Tuan A. Nguyen, Subbarao Kambhampati

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

4 Citations (Scopus)

Abstract

Most current planners assume complete domain models and focus on generating correct plans. Unfortunately, domain modeling is a laborious and error-prone task, thus real world agents have to plan with incomplete domain models. While domain experts cannot guarantee completeness, often they are able to circumscribe the incompleteness of the model by providing annotations as to which parts of the domain model may be incomplete. In this paper, we study planning problems with incomplete STRIPS domain models where the annotations specify possible preconditions and effects of actions. We show that the problem of assessing the quality of a plan, or its plan robustness, is #P-complete, establishing its equivalence with the weighted model counting problems. We introduce two approximations, lower and upper bound, for plan robustness, and then utilize them to derive heuristics for synthesizing robust plans. Our planning system, PISA, incorporating stochastic local search with these novel techniques outperforms a state-of-the-art planner handling incomplete domains in most of the tested domains, both in terms of plan quality and planning time.

Original languageEnglish (US)
Title of host publicationProceedings International Conference on Automated Planning and Scheduling, ICAPS
PublisherAAAI press
Pages181-189
Number of pages9
Volume2014-January
EditionJanuary
StatePublished - 2014
Event24th International Conference on Automated Planning and Scheduling, ICAPS 2014 - Portsmouth, United States
Duration: Jun 21 2014Jun 26 2014

Other

Other24th International Conference on Automated Planning and Scheduling, ICAPS 2014
CountryUnited States
CityPortsmouth
Period6/21/146/26/14

Fingerprint

Planning
Heuristics
Robustness
Annotation

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Information Systems and Management

Cite this

Nguyen, T. A., & Kambhampati, S. (2014). A heuristic approach to planning with incomplete STRIPS action models. In Proceedings International Conference on Automated Planning and Scheduling, ICAPS (January ed., Vol. 2014-January, pp. 181-189). AAAI press.

A heuristic approach to planning with incomplete STRIPS action models. / Nguyen, Tuan A.; Kambhampati, Subbarao.

Proceedings International Conference on Automated Planning and Scheduling, ICAPS. Vol. 2014-January January. ed. AAAI press, 2014. p. 181-189.

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

Nguyen, TA & Kambhampati, S 2014, A heuristic approach to planning with incomplete STRIPS action models. in Proceedings International Conference on Automated Planning and Scheduling, ICAPS. January edn, vol. 2014-January, AAAI press, pp. 181-189, 24th International Conference on Automated Planning and Scheduling, ICAPS 2014, Portsmouth, United States, 6/21/14.
Nguyen TA, Kambhampati S. A heuristic approach to planning with incomplete STRIPS action models. In Proceedings International Conference on Automated Planning and Scheduling, ICAPS. January ed. Vol. 2014-January. AAAI press. 2014. p. 181-189
Nguyen, Tuan A. ; Kambhampati, Subbarao. / A heuristic approach to planning with incomplete STRIPS action models. Proceedings International Conference on Automated Planning and Scheduling, ICAPS. Vol. 2014-January January. ed. AAAI press, 2014. pp. 181-189
@inproceedings{2288d312d202486caf295fe3837dfcdd,
title = "A heuristic approach to planning with incomplete STRIPS action models",
abstract = "Most current planners assume complete domain models and focus on generating correct plans. Unfortunately, domain modeling is a laborious and error-prone task, thus real world agents have to plan with incomplete domain models. While domain experts cannot guarantee completeness, often they are able to circumscribe the incompleteness of the model by providing annotations as to which parts of the domain model may be incomplete. In this paper, we study planning problems with incomplete STRIPS domain models where the annotations specify possible preconditions and effects of actions. We show that the problem of assessing the quality of a plan, or its plan robustness, is #P-complete, establishing its equivalence with the weighted model counting problems. We introduce two approximations, lower and upper bound, for plan robustness, and then utilize them to derive heuristics for synthesizing robust plans. Our planning system, PISA, incorporating stochastic local search with these novel techniques outperforms a state-of-the-art planner handling incomplete domains in most of the tested domains, both in terms of plan quality and planning time.",
author = "Nguyen, {Tuan A.} and Subbarao Kambhampati",
year = "2014",
language = "English (US)",
volume = "2014-January",
pages = "181--189",
booktitle = "Proceedings International Conference on Automated Planning and Scheduling, ICAPS",
publisher = "AAAI press",
edition = "January",

}

TY - GEN

T1 - A heuristic approach to planning with incomplete STRIPS action models

AU - Nguyen, Tuan A.

AU - Kambhampati, Subbarao

PY - 2014

Y1 - 2014

N2 - Most current planners assume complete domain models and focus on generating correct plans. Unfortunately, domain modeling is a laborious and error-prone task, thus real world agents have to plan with incomplete domain models. While domain experts cannot guarantee completeness, often they are able to circumscribe the incompleteness of the model by providing annotations as to which parts of the domain model may be incomplete. In this paper, we study planning problems with incomplete STRIPS domain models where the annotations specify possible preconditions and effects of actions. We show that the problem of assessing the quality of a plan, or its plan robustness, is #P-complete, establishing its equivalence with the weighted model counting problems. We introduce two approximations, lower and upper bound, for plan robustness, and then utilize them to derive heuristics for synthesizing robust plans. Our planning system, PISA, incorporating stochastic local search with these novel techniques outperforms a state-of-the-art planner handling incomplete domains in most of the tested domains, both in terms of plan quality and planning time.

AB - Most current planners assume complete domain models and focus on generating correct plans. Unfortunately, domain modeling is a laborious and error-prone task, thus real world agents have to plan with incomplete domain models. While domain experts cannot guarantee completeness, often they are able to circumscribe the incompleteness of the model by providing annotations as to which parts of the domain model may be incomplete. In this paper, we study planning problems with incomplete STRIPS domain models where the annotations specify possible preconditions and effects of actions. We show that the problem of assessing the quality of a plan, or its plan robustness, is #P-complete, establishing its equivalence with the weighted model counting problems. We introduce two approximations, lower and upper bound, for plan robustness, and then utilize them to derive heuristics for synthesizing robust plans. Our planning system, PISA, incorporating stochastic local search with these novel techniques outperforms a state-of-the-art planner handling incomplete domains in most of the tested domains, both in terms of plan quality and planning time.

UR - http://www.scopus.com/inward/record.url?scp=84933036236&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84933036236&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84933036236

VL - 2014-January

SP - 181

EP - 189

BT - Proceedings International Conference on Automated Planning and Scheduling, ICAPS

PB - AAAI press

ER -