Planning through Automatic Portfolio Configuration

The PbP Approach

Alfonso E. Gerevini, Alessandro Saetti, Mauro Vallati

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

In the field of domain-independent planning, several powerful planners implementing different techniques have been developed. However, no one of these systems outperforms all others in every known benchmark domain. In this work, we propose a multi-planner approach that automatically configures a portfolio of planning techniques for each given domain. The configuration process for a given domain uses a set of training instances to: (i) compute and analyze some alternative sets of macro-actions for each planner in the portfolio identifying a (possibly empty) useful set, (ii) select a cluster of planners, each one with the identified useful set of macro-actions, that is expected to perform best, and (iii) derive some additional information for configuring the execution scheduling of the selected planners at planning time. The resulting planning system, called PbP (Portfolio- based Planner), has two variants focusing on speed and plan quality. Different versions of PbP entered and won the learning track of the sixth and seventh International Planning Competitions. In this paper, we experimentally analyze PbP considering planning speed and plan quality in depth. We provide a collection of results that help to understand PbPs behavior, and demonstrate the effectiveness of our approach to configuring a portfolio of planners with macro-actions.
Original languageEnglish
Pages (from-to)639-696
Number of pages58
JournalJournal of Artificial Intelligence Research
Volume50
DOIs
Publication statusPublished - Jul 2014

Fingerprint

Planning
Macros
Scheduling

Cite this

@article{0607a0689a254eada6605e21e4a881b8,
title = "Planning through Automatic Portfolio Configuration: The PbP Approach",
abstract = "In the field of domain-independent planning, several powerful planners implementing different techniques have been developed. However, no one of these systems outperforms all others in every known benchmark domain. In this work, we propose a multi-planner approach that automatically configures a portfolio of planning techniques for each given domain. The configuration process for a given domain uses a set of training instances to: (i) compute and analyze some alternative sets of macro-actions for each planner in the portfolio identifying a (possibly empty) useful set, (ii) select a cluster of planners, each one with the identified useful set of macro-actions, that is expected to perform best, and (iii) derive some additional information for configuring the execution scheduling of the selected planners at planning time. The resulting planning system, called PbP (Portfolio- based Planner), has two variants focusing on speed and plan quality. Different versions of PbP entered and won the learning track of the sixth and seventh International Planning Competitions. In this paper, we experimentally analyze PbP considering planning speed and plan quality in depth. We provide a collection of results that help to understand PbPs behavior, and demonstrate the effectiveness of our approach to configuring a portfolio of planners with macro-actions.",
keywords = "Automated Planning, portfolio-based planner",
author = "Gerevini, {Alfonso E.} and Alessandro Saetti and Mauro Vallati",
year = "2014",
month = "7",
doi = "10.1613/jair.4359",
language = "English",
volume = "50",
pages = "639--696",
journal = "Journal of Artificial Intelligence Research",
issn = "1076-9757",
publisher = "Morgan Kaufmann Publishers, Inc.",

}

Planning through Automatic Portfolio Configuration : The PbP Approach. / Gerevini, Alfonso E.; Saetti, Alessandro; Vallati, Mauro.

In: Journal of Artificial Intelligence Research, Vol. 50, 07.2014, p. 639-696.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Planning through Automatic Portfolio Configuration

T2 - The PbP Approach

AU - Gerevini, Alfonso E.

AU - Saetti, Alessandro

AU - Vallati, Mauro

PY - 2014/7

Y1 - 2014/7

N2 - In the field of domain-independent planning, several powerful planners implementing different techniques have been developed. However, no one of these systems outperforms all others in every known benchmark domain. In this work, we propose a multi-planner approach that automatically configures a portfolio of planning techniques for each given domain. The configuration process for a given domain uses a set of training instances to: (i) compute and analyze some alternative sets of macro-actions for each planner in the portfolio identifying a (possibly empty) useful set, (ii) select a cluster of planners, each one with the identified useful set of macro-actions, that is expected to perform best, and (iii) derive some additional information for configuring the execution scheduling of the selected planners at planning time. The resulting planning system, called PbP (Portfolio- based Planner), has two variants focusing on speed and plan quality. Different versions of PbP entered and won the learning track of the sixth and seventh International Planning Competitions. In this paper, we experimentally analyze PbP considering planning speed and plan quality in depth. We provide a collection of results that help to understand PbPs behavior, and demonstrate the effectiveness of our approach to configuring a portfolio of planners with macro-actions.

AB - In the field of domain-independent planning, several powerful planners implementing different techniques have been developed. However, no one of these systems outperforms all others in every known benchmark domain. In this work, we propose a multi-planner approach that automatically configures a portfolio of planning techniques for each given domain. The configuration process for a given domain uses a set of training instances to: (i) compute and analyze some alternative sets of macro-actions for each planner in the portfolio identifying a (possibly empty) useful set, (ii) select a cluster of planners, each one with the identified useful set of macro-actions, that is expected to perform best, and (iii) derive some additional information for configuring the execution scheduling of the selected planners at planning time. The resulting planning system, called PbP (Portfolio- based Planner), has two variants focusing on speed and plan quality. Different versions of PbP entered and won the learning track of the sixth and seventh International Planning Competitions. In this paper, we experimentally analyze PbP considering planning speed and plan quality in depth. We provide a collection of results that help to understand PbPs behavior, and demonstrate the effectiveness of our approach to configuring a portfolio of planners with macro-actions.

KW - Automated Planning

KW - portfolio-based planner

U2 - 10.1613/jair.4359

DO - 10.1613/jair.4359

M3 - Article

VL - 50

SP - 639

EP - 696

JO - Journal of Artificial Intelligence Research

JF - Journal of Artificial Intelligence Research

SN - 1076-9757

ER -