Exploring the Synergy between Two Modular Learning Techniques for Automated Planning

Raquel Fuentetaja, Lukáš Chrpa, Thomas L. McCluskey, Mauro Vallati

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

Abstract

In the last decade the emphasis on improving the operational performance of domain independent automated planners has been in developing complex techniques which merge a range of different strategies. This quest for operational advantage, driven by the regular international planning competitions, has not made it easy to study, understand and predict what combinations of techniques will have what effect on a planner’s behaviour in a particular application domain. In this paper, we consider two machine learning techniques for planner performance improvement, and exploit a modular approach to their combination in order to facilitate the analysis of the impact of each individual component. We believe this can contribute to the development of more transparent planning engines, which are designed using modular, interchangeable, and well-founded components. Specifically, we combined two previously unrelated learning techniques, entanglements and relational decision trees, to guide a “vanilla” search algorithm. We report on a large experimental analysis which demonstrates the effectiveness of the approach in terms of performance improvements, resulting in a very competitive planning configuration despite the use of a more modular and transparent architecture. This gives insights on the strengths and weaknesses of the considered approaches, that will help their future exploitation.

LanguageEnglish
Title of host publicationProceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015
PublisherAAAI press
Pages35-43
Number of pages9
ISBN (Electronic)9781577357322
Publication statusPublished - 13 May 2015
Event8th Annual Symposium on Combinatorial Search - Ein Gedi, Israel
Duration: 11 Jun 201513 Jun 2015
Conference number: 8
http://www.ise.bgu.ac.il/socs2015/ (Link to Conference Website)

Conference

Conference8th Annual Symposium on Combinatorial Search
Abbreviated titleSoCS 2015
CountryIsrael
CityEin Gedi
Period11/06/1513/06/15
Internet address

Fingerprint

Planning
Decision trees
Learning systems
Engines

Cite this

Fuentetaja, R., Chrpa, L., McCluskey, T. L., & Vallati, M. (2015). Exploring the Synergy between Two Modular Learning Techniques for Automated Planning. In Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015 (pp. 35-43). AAAI press.
Fuentetaja, Raquel ; Chrpa, Lukáš ; McCluskey, Thomas L. ; Vallati, Mauro. / Exploring the Synergy between Two Modular Learning Techniques for Automated Planning. Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015. AAAI press, 2015. pp. 35-43
@inproceedings{ef5b88754bdc4f7caddc4a33c82bede8,
title = "Exploring the Synergy between Two Modular Learning Techniques for Automated Planning",
abstract = "In the last decade the emphasis on improving the operational performance of domain independent automated planners has been in developing complex techniques which merge a range of different strategies. This quest for operational advantage, driven by the regular international planning competitions, has not made it easy to study, understand and predict what combinations of techniques will have what effect on a planner’s behaviour in a particular application domain. In this paper, we consider two machine learning techniques for planner performance improvement, and exploit a modular approach to their combination in order to facilitate the analysis of the impact of each individual component. We believe this can contribute to the development of more transparent planning engines, which are designed using modular, interchangeable, and well-founded components. Specifically, we combined two previously unrelated learning techniques, entanglements and relational decision trees, to guide a “vanilla” search algorithm. We report on a large experimental analysis which demonstrates the effectiveness of the approach in terms of performance improvements, resulting in a very competitive planning configuration despite the use of a more modular and transparent architecture. This gives insights on the strengths and weaknesses of the considered approaches, that will help their future exploitation.",
keywords = "Classical Planning, Learning for Planning, Entanglements, Relational decision trees",
author = "Raquel Fuentetaja and Luk{\'a}š Chrpa and McCluskey, {Thomas L.} and Mauro Vallati",
year = "2015",
month = "5",
day = "13",
language = "English",
pages = "35--43",
booktitle = "Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015",
publisher = "AAAI press",

}

Fuentetaja, R, Chrpa, L, McCluskey, TL & Vallati, M 2015, Exploring the Synergy between Two Modular Learning Techniques for Automated Planning. in Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015. AAAI press, pp. 35-43, 8th Annual Symposium on Combinatorial Search, Ein Gedi, Israel, 11/06/15.

Exploring the Synergy between Two Modular Learning Techniques for Automated Planning. / Fuentetaja, Raquel; Chrpa, Lukáš; McCluskey, Thomas L.; Vallati, Mauro.

Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015. AAAI press, 2015. p. 35-43.

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

TY - GEN

T1 - Exploring the Synergy between Two Modular Learning Techniques for Automated Planning

AU - Fuentetaja, Raquel

AU - Chrpa, Lukáš

AU - McCluskey, Thomas L.

AU - Vallati, Mauro

PY - 2015/5/13

Y1 - 2015/5/13

N2 - In the last decade the emphasis on improving the operational performance of domain independent automated planners has been in developing complex techniques which merge a range of different strategies. This quest for operational advantage, driven by the regular international planning competitions, has not made it easy to study, understand and predict what combinations of techniques will have what effect on a planner’s behaviour in a particular application domain. In this paper, we consider two machine learning techniques for planner performance improvement, and exploit a modular approach to their combination in order to facilitate the analysis of the impact of each individual component. We believe this can contribute to the development of more transparent planning engines, which are designed using modular, interchangeable, and well-founded components. Specifically, we combined two previously unrelated learning techniques, entanglements and relational decision trees, to guide a “vanilla” search algorithm. We report on a large experimental analysis which demonstrates the effectiveness of the approach in terms of performance improvements, resulting in a very competitive planning configuration despite the use of a more modular and transparent architecture. This gives insights on the strengths and weaknesses of the considered approaches, that will help their future exploitation.

AB - In the last decade the emphasis on improving the operational performance of domain independent automated planners has been in developing complex techniques which merge a range of different strategies. This quest for operational advantage, driven by the regular international planning competitions, has not made it easy to study, understand and predict what combinations of techniques will have what effect on a planner’s behaviour in a particular application domain. In this paper, we consider two machine learning techniques for planner performance improvement, and exploit a modular approach to their combination in order to facilitate the analysis of the impact of each individual component. We believe this can contribute to the development of more transparent planning engines, which are designed using modular, interchangeable, and well-founded components. Specifically, we combined two previously unrelated learning techniques, entanglements and relational decision trees, to guide a “vanilla” search algorithm. We report on a large experimental analysis which demonstrates the effectiveness of the approach in terms of performance improvements, resulting in a very competitive planning configuration despite the use of a more modular and transparent architecture. This gives insights on the strengths and weaknesses of the considered approaches, that will help their future exploitation.

KW - Classical Planning

KW - Learning for Planning

KW - Entanglements

KW - Relational decision trees

UR - http://www.scopus.com/inward/record.url?scp=85048721927&partnerID=8YFLogxK

M3 - Conference contribution

SP - 35

EP - 43

BT - Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015

PB - AAAI press

ER -

Fuentetaja R, Chrpa L, McCluskey TL, Vallati M. Exploring the Synergy between Two Modular Learning Techniques for Automated Planning. In Proceedings of the 8th Annual Symposium on Combinatorial Search, SoCS 2015. AAAI press. 2015. p. 35-43