Learning to improve agent behaviours in GOAL

Dhirendra Singh; Koen V. Hindriks

doi:10.1007/978-3-642-38700-5_10

Learning to improve agent behaviours in GOAL

Research output: Chapter in Book / Report / Conference proceeding › Conference contribution › Academic › peer-review

Abstract

This paper investigates the issue of adaptability of behaviour in the context of agent-oriented programming. We focus on improving action selection in rule-based agent programming languages using a reinforcement learning mechanism under the hood. The novelty is that learning utilises the existing mental state representation of the agent, which means that (i) the programming model is unchanged and using learning within the program becomes straightforward, and (ii) adaptive behaviours can be combined with regular behaviours in a modular way. Overall, the key to effective programming in this setting is to balance between constraining behaviour using operational knowledge, and leaving flexibility to allow for ongoing adaptation. We illustrate this using different types of programs for solving the Blocks World problem.

Original language	English
Title of host publication	Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers
Pages	158-173
Number of pages	16
DOIs	https://doi.org/10.1007/978-3-642-38700-5_10
Publication status	Published - 5 Sept 2013
Externally published	Yes
Event	10th International Workshop on Programming Multi-Agent Systems, ProMAS 2012 - Valencia, Spain Duration: 5 Jun 2012 → 5 Jun 2012

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	7837 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	10th International Workshop on Programming Multi-Agent Systems, ProMAS 2012
Country/Territory	Spain
City	Valencia
Period	5/06/12 → 5/06/12

Keywords

Agent programming
reinforcement learning
rule selection

Access to Document

10.1007/978-3-642-38700-5_10

Persistent URL (handle)

Persistent link

Cite this

Singh, D., & Hindriks, K. V. (2013). Learning to improve agent behaviours in GOAL. In Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers (pp. 158-173). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7837 LNAI). https://doi.org/10.1007/978-3-642-38700-5_10

@inproceedings{9968c7b3b0bb42dab0637dd6cadb4b57,

title = "Learning to improve agent behaviours in GOAL",

abstract = "This paper investigates the issue of adaptability of behaviour in the context of agent-oriented programming. We focus on improving action selection in rule-based agent programming languages using a reinforcement learning mechanism under the hood. The novelty is that learning utilises the existing mental state representation of the agent, which means that (i) the programming model is unchanged and using learning within the program becomes straightforward, and (ii) adaptive behaviours can be combined with regular behaviours in a modular way. Overall, the key to effective programming in this setting is to balance between constraining behaviour using operational knowledge, and leaving flexibility to allow for ongoing adaptation. We illustrate this using different types of programs for solving the Blocks World problem.",

keywords = "Agent programming, reinforcement learning, rule selection",

author = "Dhirendra Singh and Hindriks, {Koen V.}",

year = "2013",

month = sep,

day = "5",

doi = "10.1007/978-3-642-38700-5_10",

language = "English",

isbn = "9783642386992",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "158--173",

booktitle = "Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers",

note = "10th International Workshop on Programming Multi-Agent Systems, ProMAS 2012 ; Conference date: 05-06-2012 Through 05-06-2012",

}

Singh, D & Hindriks, KV 2013, Learning to improve agent behaviours in GOAL. in Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7837 LNAI, pp. 158-173, 10th International Workshop on Programming Multi-Agent Systems, ProMAS 2012, Valencia, Spain, 5/06/12. https://doi.org/10.1007/978-3-642-38700-5_10

Learning to improve agent behaviours in GOAL. / Singh, Dhirendra; Hindriks, Koen V.
Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers. 2013. p. 158-173 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7837 LNAI).

Research output: Chapter in Book / Report / Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Learning to improve agent behaviours in GOAL

AU - Singh, Dhirendra

AU - Hindriks, Koen V.

PY - 2013/9/5

Y1 - 2013/9/5

N2 - This paper investigates the issue of adaptability of behaviour in the context of agent-oriented programming. We focus on improving action selection in rule-based agent programming languages using a reinforcement learning mechanism under the hood. The novelty is that learning utilises the existing mental state representation of the agent, which means that (i) the programming model is unchanged and using learning within the program becomes straightforward, and (ii) adaptive behaviours can be combined with regular behaviours in a modular way. Overall, the key to effective programming in this setting is to balance between constraining behaviour using operational knowledge, and leaving flexibility to allow for ongoing adaptation. We illustrate this using different types of programs for solving the Blocks World problem.

AB - This paper investigates the issue of adaptability of behaviour in the context of agent-oriented programming. We focus on improving action selection in rule-based agent programming languages using a reinforcement learning mechanism under the hood. The novelty is that learning utilises the existing mental state representation of the agent, which means that (i) the programming model is unchanged and using learning within the program becomes straightforward, and (ii) adaptive behaviours can be combined with regular behaviours in a modular way. Overall, the key to effective programming in this setting is to balance between constraining behaviour using operational knowledge, and leaving flexibility to allow for ongoing adaptation. We illustrate this using different types of programs for solving the Blocks World problem.

KW - Agent programming

KW - reinforcement learning

KW - rule selection

UR - http://www.scopus.com/inward/record.url?scp=84883265063&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883265063&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-38700-5_10

DO - 10.1007/978-3-642-38700-5_10

M3 - Conference contribution

AN - SCOPUS:84883265063

SN - 9783642386992

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 158

EP - 173

BT - Programming Multi-Agent Systems - 10th International Workshop, ProMAS 2012, Revised Selected Papers

T2 - 10th International Workshop on Programming Multi-Agent Systems, ProMAS 2012

Y2 - 5 June 2012 through 5 June 2012

ER -

Learning to improve agent behaviours in GOAL

Abstract

Publication series

Conference

Keywords

Access to Document

Persistent URL (handle)

Other files and links

Fingerprint

Cite this