Toward approximate planning in very large stochastic domains

Ann E. Nicholson; Leslie Pack Kaelbling

Toward approximate planning in very large stochastic domains

Research output: Contribution to conference › Paper › peer-review

Abstract

In this paper we extend previous work on approximate planning in large stochastic domains by adding the ability to plan in automatically-generated abstract world views. The dynamics of the domain are represented compositionally using a Bayesian network. Sensitivity analysis is performed on the network to identify the aspects of the world upon which success is most highly dependent. An abstract world model is constructed by including only the most relevant aspects of the world. The world view can be refined over time, making the overall planner behave in most cases like an anytime algorithm. This paper is a preliminary report on this ongoing work.

Original language	English
Pages	190-196
Number of pages	7
Publication status	Published - 1994
Externally published	Yes
Event	AAAI Spring Symposium 1994 - Palo Alto, United States of America Duration: 21 Mar 1994 → 23 Mar 1994

Conference

Conference	AAAI Spring Symposium 1994
Country/Territory	United States of America
City	Palo Alto
Period	21/03/94 → 23/03/94

Cite this

@conference{3666d27d4bb14f6885ca337cdb0859b6,

title = "Toward approximate planning in very large stochastic domains",

abstract = "In this paper we extend previous work on approximate planning in large stochastic domains by adding the ability to plan in automatically-generated abstract world views. The dynamics of the domain are represented compositionally using a Bayesian network. Sensitivity analysis is performed on the network to identify the aspects of the world upon which success is most highly dependent. An abstract world model is constructed by including only the most relevant aspects of the world. The world view can be refined over time, making the overall planner behave in most cases like an anytime algorithm. This paper is a preliminary report on this ongoing work.",

author = "Nicholson, {Ann E.} and Kaelbling, {Leslie Pack}",

note = "Funding Information: *AnnN icholson{\textquoteright}s workw as supported by the Advanced ResearchP rojects Agencyo f the Departmenotf Defensem on-itored by the Air Force under Contract No. F30602-91-C-0041 and by the National Science foundationi n conjunction with the AdvancedR esearchP rojects Agencyo f the Depart-mento f Defense under Contract No. IRI-8905436. tLeslie Kaelbling{\textquoteright}s workw as supported in part by a National Science FoundationN ational YoungI nvestigator AwardI RI-9257592 and in part by ONRC ontract N00014-91-4052, ARPAO rder 8225. Publisher Copyright: {\textcopyright} 1994, AAAI (www.aaai.org). All rights reserved.; AAAI Spring Symposium 1994 ; Conference date: 21-03-1994 Through 23-03-1994",

year = "1994",

language = "English",

pages = "190--196",

}

TY - CONF

T1 - Toward approximate planning in very large stochastic domains

AU - Nicholson, Ann E.

AU - Kaelbling, Leslie Pack

N1 - Funding Information: *AnnN icholson’s workw as supported by the Advanced ResearchP rojects Agencyo f the Departmenotf Defensem on-itored by the Air Force under Contract No. F30602-91-C-0041 and by the National Science foundationi n conjunction with the AdvancedR esearchP rojects Agencyo f the Depart-mento f Defense under Contract No. IRI-8905436. tLeslie Kaelbling’s workw as supported in part by a National Science FoundationN ational YoungI nvestigator AwardI RI-9257592 and in part by ONRC ontract N00014-91-4052, ARPAO rder 8225. Publisher Copyright: © 1994, AAAI (www.aaai.org). All rights reserved.

PY - 1994

Y1 - 1994

N2 - In this paper we extend previous work on approximate planning in large stochastic domains by adding the ability to plan in automatically-generated abstract world views. The dynamics of the domain are represented compositionally using a Bayesian network. Sensitivity analysis is performed on the network to identify the aspects of the world upon which success is most highly dependent. An abstract world model is constructed by including only the most relevant aspects of the world. The world view can be refined over time, making the overall planner behave in most cases like an anytime algorithm. This paper is a preliminary report on this ongoing work.

AB - In this paper we extend previous work on approximate planning in large stochastic domains by adding the ability to plan in automatically-generated abstract world views. The dynamics of the domain are represented compositionally using a Bayesian network. Sensitivity analysis is performed on the network to identify the aspects of the world upon which success is most highly dependent. An abstract world model is constructed by including only the most relevant aspects of the world. The world view can be refined over time, making the overall planner behave in most cases like an anytime algorithm. This paper is a preliminary report on this ongoing work.

UR - http://www.scopus.com/inward/record.url?scp=85169001561&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85169001561

SP - 190

EP - 196

T2 - AAAI Spring Symposium 1994

Y2 - 21 March 1994 through 23 March 1994

ER -

Toward approximate planning in very large stochastic domains

Abstract

Conference

Other files and links

Cite this