Static and Dynamic Portfolio Methods for Optimal Planning: An Empirical Analysis

Mattia Rizzini, Chris Fawcett, Mauro Vallati, Alfonso E. Gerevini, Holger H. Hoos

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning.

Here, we consider the construction of sequential planner portfolios for domainindependent optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive empirical analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.
LanguageEnglish
Article number1760006
Number of pages27
JournalInternational Journal on Artificial Intelligence Tools
Volume26
Issue number01
DOIs
Publication statusPublished - 23 Feb 2017

Fingerprint

Planning

Cite this

Rizzini, Mattia ; Fawcett, Chris ; Vallati, Mauro ; Gerevini, Alfonso E. ; Hoos, Holger H. / Static and Dynamic Portfolio Methods for Optimal Planning : An Empirical Analysis. In: International Journal on Artificial Intelligence Tools. 2017 ; Vol. 26, No. 01.
@article{f3713b9b1d3d44a395e1e9415031bf1f,
title = "Static and Dynamic Portfolio Methods for Optimal Planning: An Empirical Analysis",
abstract = "Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning.Here, we consider the construction of sequential planner portfolios for domainindependent optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive empirical analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.",
keywords = "Automated Planning, optimal planning, sequential portfolio, per-instance portfolio generation",
author = "Mattia Rizzini and Chris Fawcett and Mauro Vallati and Gerevini, {Alfonso E.} and Hoos, {Holger H.}",
year = "2017",
month = "2",
day = "23",
doi = "10.1142/S0218213017600065",
language = "English",
volume = "26",
journal = "International Journal on Artificial Intelligence Tools",
issn = "0218-2130",
publisher = "World Scientific Publishing Co. Pte Ltd",
number = "01",

}

Static and Dynamic Portfolio Methods for Optimal Planning : An Empirical Analysis. / Rizzini, Mattia; Fawcett, Chris; Vallati, Mauro; Gerevini, Alfonso E.; Hoos, Holger H.

In: International Journal on Artificial Intelligence Tools, Vol. 26, No. 01, 1760006, 23.02.2017.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Static and Dynamic Portfolio Methods for Optimal Planning

T2 - International Journal on Artificial Intelligence Tools

AU - Rizzini, Mattia

AU - Fawcett, Chris

AU - Vallati, Mauro

AU - Gerevini, Alfonso E.

AU - Hoos, Holger H.

PY - 2017/2/23

Y1 - 2017/2/23

N2 - Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning.Here, we consider the construction of sequential planner portfolios for domainindependent optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive empirical analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.

AB - Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning.Here, we consider the construction of sequential planner portfolios for domainindependent optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive empirical analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.

KW - Automated Planning

KW - optimal planning

KW - sequential portfolio

KW - per-instance portfolio generation

U2 - 10.1142/S0218213017600065

DO - 10.1142/S0218213017600065

M3 - Article

VL - 26

JO - International Journal on Artificial Intelligence Tools

JF - International Journal on Artificial Intelligence Tools

SN - 0218-2130

IS - 01

M1 - 1760006

ER -