Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis

Raquel Fuentetaja, Michael W Barley, Daniel Borrajo, Jordan Douglas, Santiago Franco Aixela, Patricia J Riddle

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

Abstract

Two key aspects of problem solving are representation and search heuristics. Both theoretical and experimental studies have shown that there is no one best problem representation nor one best search heuristic. Therefore, some recent methods, e.g., portfolios, learn a good combination of problem solvers to be used in a given domain or set of domains. There are even dynamic portfolios that select a particular combination of problem solvers specific to a problem. These approaches: (1) need to perform a learning step; (2) do not usually focus on changing the representation of the input domain/problem; and (3) frequently do not adapt the portfolio to the specific problem. This paper describes a meta-reasoning system that searches through the space of combinations of representations and heuristics to find one suitable for optimally solving the specific problem. We show that this approach can be better than selecting
a combination to use for all problems within a domain and is competitive with state of the art optimal planners.
LanguageEnglish
Title of host publicationProceedings of the Thirty-Second AAAI Conference on Artificial Intelligence
PublisherAAAI press
Pages6169-6176
Number of pages8
Publication statusPublished - 26 Apr 2018
Event32nd Association for the Advancement of Artificial Intelligence Conference - Hilton New Orleans Riverside, New Orleans, United States
Duration: 2 Feb 20187 Feb 2018
Conference number: 32
https://aaai.org/Conferences/AAAI-18/ (Link to Conference Details)

Conference

Conference32nd Association for the Advancement of Artificial Intelligence Conference
Abbreviated titleAAAI-18
CountryUnited States
CityNew Orleans
Period2/02/187/02/18
Internet address

Cite this

Fuentetaja, R., Barley, M. W., Borrajo, D., Douglas, J., Franco Aixela, S., & Riddle, P. J. (2018). Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence (pp. 6169-6176). AAAI press.
Fuentetaja, Raquel ; Barley, Michael W ; Borrajo, Daniel ; Douglas, Jordan ; Franco Aixela, Santiago ; Riddle, Patricia J. / Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis. Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence. AAAI press, 2018. pp. 6169-6176
@inproceedings{fa0d173577364dd68efa6ebaed2887b9,
title = "Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis",
abstract = "Two key aspects of problem solving are representation and search heuristics. Both theoretical and experimental studies have shown that there is no one best problem representation nor one best search heuristic. Therefore, some recent methods, e.g., portfolios, learn a good combination of problem solvers to be used in a given domain or set of domains. There are even dynamic portfolios that select a particular combination of problem solvers specific to a problem. These approaches: (1) need to perform a learning step; (2) do not usually focus on changing the representation of the input domain/problem; and (3) frequently do not adapt the portfolio to the specific problem. This paper describes a meta-reasoning system that searches through the space of combinations of representations and heuristics to find one suitable for optimally solving the specific problem. We show that this approach can be better than selectinga combination to use for all problems within a domain and is competitive with state of the art optimal planners.",
keywords = "Planning, meta-reasoning, Search algorithms, reformulation, heuristic selection, representation selection, Predictive modelling",
author = "Raquel Fuentetaja and Barley, {Michael W} and Daniel Borrajo and Jordan Douglas and {Franco Aixela}, Santiago and Riddle, {Patricia J}",
year = "2018",
month = "4",
day = "26",
language = "English",
pages = "6169--6176",
booktitle = "Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence",
publisher = "AAAI press",

}

Fuentetaja, R, Barley, MW, Borrajo, D, Douglas, J, Franco Aixela, S & Riddle, PJ 2018, Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis. in Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence. AAAI press, pp. 6169-6176, 32nd Association for the Advancement of Artificial Intelligence Conference , New Orleans, United States, 2/02/18.

Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis. / Fuentetaja, Raquel; Barley, Michael W; Borrajo, Daniel; Douglas, Jordan; Franco Aixela, Santiago; Riddle, Patricia J.

Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence. AAAI press, 2018. p. 6169-6176.

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

TY - GEN

T1 - Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis

AU - Fuentetaja, Raquel

AU - Barley, Michael W

AU - Borrajo, Daniel

AU - Douglas, Jordan

AU - Franco Aixela, Santiago

AU - Riddle, Patricia J

PY - 2018/4/26

Y1 - 2018/4/26

N2 - Two key aspects of problem solving are representation and search heuristics. Both theoretical and experimental studies have shown that there is no one best problem representation nor one best search heuristic. Therefore, some recent methods, e.g., portfolios, learn a good combination of problem solvers to be used in a given domain or set of domains. There are even dynamic portfolios that select a particular combination of problem solvers specific to a problem. These approaches: (1) need to perform a learning step; (2) do not usually focus on changing the representation of the input domain/problem; and (3) frequently do not adapt the portfolio to the specific problem. This paper describes a meta-reasoning system that searches through the space of combinations of representations and heuristics to find one suitable for optimally solving the specific problem. We show that this approach can be better than selectinga combination to use for all problems within a domain and is competitive with state of the art optimal planners.

AB - Two key aspects of problem solving are representation and search heuristics. Both theoretical and experimental studies have shown that there is no one best problem representation nor one best search heuristic. Therefore, some recent methods, e.g., portfolios, learn a good combination of problem solvers to be used in a given domain or set of domains. There are even dynamic portfolios that select a particular combination of problem solvers specific to a problem. These approaches: (1) need to perform a learning step; (2) do not usually focus on changing the representation of the input domain/problem; and (3) frequently do not adapt the portfolio to the specific problem. This paper describes a meta-reasoning system that searches through the space of combinations of representations and heuristics to find one suitable for optimally solving the specific problem. We show that this approach can be better than selectinga combination to use for all problems within a domain and is competitive with state of the art optimal planners.

KW - Planning

KW - meta-reasoning

KW - Search algorithms

KW - reformulation

KW - heuristic selection

KW - representation selection

KW - Predictive modelling

M3 - Conference contribution

SP - 6169

EP - 6176

BT - Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence

PB - AAAI press

ER -

Fuentetaja R, Barley MW, Borrajo D, Douglas J, Franco Aixela S, Riddle PJ. Meta-Search Through the Space of Representations and Heuristics on a Problem by Problem Basis. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence. AAAI press. 2018. p. 6169-6176