Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability

Alfonso Emilio Gerevini, Alessandro Saetti, Mauro Vallati

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

4 Citations (Scopus)

Abstract

The use of automatically learned knowledge for a planning domain can significantly improve the performance of a generic planner when solving a problem in this domain. In this work, we focus on the well-known SAT-based approach to planning and investigate two types of learned knowledge that have not been studied in this planning framework before: macro-actions and planning horizon. Macro-actions are sequences of actions that typically occur in the solution plans, while a planning horizon of a problem is the length of a (possibly optimal) plan solving it. We propose a method that uses a machine learning tool for building a predictive model of the optimal planning horizon, and variants of the well-known planner and solver that can exploit macro actions and learned planning horizons to improve their performance. An experimental analysis illustrates the effectiveness of the proposed techniques.

LanguageEnglish
Title of host publicationAI*IA 2011
Subtitle of host publicationArtificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings
EditorsRoberto Pirrone, Filippo Sorbello
PublisherSpringer Verlag
Pages189-200
Number of pages12
ISBN (Electronic)9783642239540
ISBN (Print)9783642239533
DOIs
Publication statusPublished - 26 Sep 2011
Externally publishedYes
Event12th International Conference of the Italian Association for Artificial Intelligence: Artificial Intelligence Around Man and Beyond - Palermo, Italy
Duration: 15 Sep 201117 Sep 2011
Conference number: 12
http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=14579 (Link to Conference Website)

Publication series

NameLecture Notes in Computer Science
Volume6934
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Conference of the Italian Association for Artificial Intelligence
Abbreviated titleAI*IA 2011
CountryItaly
CityPalermo
Period15/09/1117/09/11
Internet address

Fingerprint

Macros
Planning
Horizon
Predictive Model
Experimental Analysis
Learning systems
Machine Learning

Cite this

Gerevini, A. E., Saetti, A., & Vallati, M. (2011). Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability. In R. Pirrone, & F. Sorbello (Eds.), AI*IA 2011: Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings (pp. 189-200). (Lecture Notes in Computer Science; Vol. 6934). Springer Verlag. https://doi.org/10.1007/978-3-642-23954-0_19
Gerevini, Alfonso Emilio ; Saetti, Alessandro ; Vallati, Mauro. / Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability. AI*IA 2011: Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings. editor / Roberto Pirrone ; Filippo Sorbello. Springer Verlag, 2011. pp. 189-200 (Lecture Notes in Computer Science).
@inproceedings{7861787edd8042e7adbd3158ad82a9e3,
title = "Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability",
abstract = "The use of automatically learned knowledge for a planning domain can significantly improve the performance of a generic planner when solving a problem in this domain. In this work, we focus on the well-known SAT-based approach to planning and investigate two types of learned knowledge that have not been studied in this planning framework before: macro-actions and planning horizon. Macro-actions are sequences of actions that typically occur in the solution plans, while a planning horizon of a problem is the length of a (possibly optimal) plan solving it. We propose a method that uses a machine learning tool for building a predictive model of the optimal planning horizon, and variants of the well-known planner and solver that can exploit macro actions and learned planning horizons to improve their performance. An experimental analysis illustrates the effectiveness of the proposed techniques.",
keywords = "Machine learning for planning, Planning as satisfiability",
author = "Gerevini, {Alfonso Emilio} and Alessandro Saetti and Mauro Vallati",
year = "2011",
month = "9",
day = "26",
doi = "10.1007/978-3-642-23954-0_19",
language = "English",
isbn = "9783642239533",
series = "Lecture Notes in Computer Science",
publisher = "Springer Verlag",
pages = "189--200",
editor = "Roberto Pirrone and Filippo Sorbello",
booktitle = "AI*IA 2011",

}

Gerevini, AE, Saetti, A & Vallati, M 2011, Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability. in R Pirrone & F Sorbello (eds), AI*IA 2011: Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings. Lecture Notes in Computer Science, vol. 6934, Springer Verlag, pp. 189-200, 12th International Conference of the Italian Association for Artificial Intelligence, Palermo, Italy, 15/09/11. https://doi.org/10.1007/978-3-642-23954-0_19

Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability. / Gerevini, Alfonso Emilio; Saetti, Alessandro; Vallati, Mauro.

AI*IA 2011: Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings. ed. / Roberto Pirrone; Filippo Sorbello. Springer Verlag, 2011. p. 189-200 (Lecture Notes in Computer Science; Vol. 6934).

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

TY - GEN

T1 - Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability

AU - Gerevini, Alfonso Emilio

AU - Saetti, Alessandro

AU - Vallati, Mauro

PY - 2011/9/26

Y1 - 2011/9/26

N2 - The use of automatically learned knowledge for a planning domain can significantly improve the performance of a generic planner when solving a problem in this domain. In this work, we focus on the well-known SAT-based approach to planning and investigate two types of learned knowledge that have not been studied in this planning framework before: macro-actions and planning horizon. Macro-actions are sequences of actions that typically occur in the solution plans, while a planning horizon of a problem is the length of a (possibly optimal) plan solving it. We propose a method that uses a machine learning tool for building a predictive model of the optimal planning horizon, and variants of the well-known planner and solver that can exploit macro actions and learned planning horizons to improve their performance. An experimental analysis illustrates the effectiveness of the proposed techniques.

AB - The use of automatically learned knowledge for a planning domain can significantly improve the performance of a generic planner when solving a problem in this domain. In this work, we focus on the well-known SAT-based approach to planning and investigate two types of learned knowledge that have not been studied in this planning framework before: macro-actions and planning horizon. Macro-actions are sequences of actions that typically occur in the solution plans, while a planning horizon of a problem is the length of a (possibly optimal) plan solving it. We propose a method that uses a machine learning tool for building a predictive model of the optimal planning horizon, and variants of the well-known planner and solver that can exploit macro actions and learned planning horizons to improve their performance. An experimental analysis illustrates the effectiveness of the proposed techniques.

KW - Machine learning for planning

KW - Planning as satisfiability

UR - http://www.scopus.com/inward/record.url?scp=80053030892&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-23954-0_19

DO - 10.1007/978-3-642-23954-0_19

M3 - Conference contribution

SN - 9783642239533

T3 - Lecture Notes in Computer Science

SP - 189

EP - 200

BT - AI*IA 2011

A2 - Pirrone, Roberto

A2 - Sorbello, Filippo

PB - Springer Verlag

ER -

Gerevini AE, Saetti A, Vallati M. Exploiting Macro-actions and Predicting Plan Length in Planning as Satisfiability. In Pirrone R, Sorbello F, editors, AI*IA 2011: Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Association for Artificial Intelligence, Proceedings. Springer Verlag. 2011. p. 189-200. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-642-23954-0_19