Intelligent Agents VII Agent Theories Architectures and Languages: 7th International Workshop, ATAL 2000 Boston, MA, USA, July 7–9, 2000 Proceedings

371 Pages·2001·2.55 MB·English

by  Michael Wooldridge

Lecture Notes in Artificial Intelligence 1986
SubseriesofLectureNotesinComputerScience
EditedbyJ.G.CarbonellandJ.Siekmann
Lecture Notes in Computer Science
EditedbyG.Goos,J.HartmanisandJ.vanLeeuwen
3
Berlin
Heidelberg
NewYork
Barcelona
HongKong
London
Milan
Paris
Singapore
Tokyo
CristianoCastelfranchi YvesLespe´rance(Eds.)
Intelligent
Agents VII
Agent Theories Architectures
and Languages
7th International Workshop, ATAL 2000
Boston, MA, USA, July 7-9, 2000
Proceedings
1 3
SeriesEditors
JaimeG.Carbonell,CarnegieMellonUniversity,Pittsburgh,PA,USA
Jo¨rgSiekmann,UniversityofSaarland,Saarbru¨cken,Germany
VolumeEditors
CristianoCastelfranchi
UniversityofSiena
DepartmentofCommunicationSciences,"CognitiveScience"
piazzaS.Francesco8,53100Siena,Italy
E-mail:[email protected]
YvesLespe´rance
YorkUniversity,DepartmentofComputerScience
4700KeeleStreet,Toronto,Onatrio,CanadaM3J1P3
E-mail:[email protected]
Cataloging-in-PublicationDataappliedfor
DieDeutscheBibliothek-CIP-Einheitsaufnahme
IntelligentagentsVII:agenttheories,architectures,andlanguages;7th
internationalworkshop;proceedings/ATAL2000,Boston,MA,USA,July7-
9,2000.ChristianoCastelfranchi;YvesLespérance(ed.).-Berlin;
Heidelberg;NewYork;Barcelona;HongKong;London;Milan;Paris;
Singapore;Tokyo:Springer,2001
(Lecturenotesincomputerscience;Vol.1986:Lecturenotesin
artificialintelligence)
ISBN3-540-42422-9
CRSubjectClassification(1998):I.2.11,I.2,C.2.4,D.2,F.3
ISBN3-540-42422-9Springer-VerlagBerlinHeidelbergNewYork
Thisworkissubjecttocopyright.Allrightsarereserved,whetherthewholeorpartofthematerialis
concerned,specificallytherightsoftranslation,reprinting,re-useofillustrations,recitation,broadcasting,
reproductiononmicrofilmsorinanyotherway,andstorageindatabanks.Duplicationofthispublication
orpartsthereofispermittedonlyundertheprovisionsoftheGermanCopyrightLawofSeptember9,1965,
initscurrentversion,andpermissionforusemustalwaysbeobtainedfromSpringer-Verlag.Violationsare
liableforprosecutionundertheGermanCopyrightLaw.
Springer-VerlagBerlinHeidelbergNewYork
amemberofBertelsmannSpringerScience+BusinessMediaGmbH
http://www.springer.de
©Springer-VerlagBerlinHeidelberg2001
PrintedinGermany
Typesetting:Camera-readybyauthor,dataconversionPTPBerlin,StefanSossna
Printedonacid-freepaper SPIN10781933 06/3142 543210
Preface
Intelligent agents are one of the most important developments in computer science of
the past decade. Agents are of interest in many important application areas, ranging
from human-computer interaction to industrial process control. TheATAL workshop
series aims to bring together researchers interested in the core/micro aspects of agent
technology. Specifically,ATAL addresses issues such as theories of agency, software
architecturesforintelligentagents,methodologiesandprogramminglanguagesforrea-
lizingagents,andsoftwaretoolsforapplyingandevaluatingagentsystems.Oneofthe
strengthsoftheATALworkshopseriesisitsemphasisonthesynergiesbetweentheories,
languages,architectures,infrastructures,methodologies,andformalmethods.
This year’s workshop continued the ATAL trend of attracting a large number of
highqualitysubmissions.Inmoredetail,71papersweresubmittedtotheATAL2000
workshop, from 21 countries.After stringent reviewing, 22 papers were accepted for
publicationandappearintheseproceedings.
Aswithpreviousworkshopsintheseries,wechosetoemphasizewhatweperceive
asimportantnewthemesinagentresearch.Thisyear’sthemeswerebothassociatedwith
thefactthatthetechnologyofintelligentagentsandmulti-agentsystemsisbeginning
tomigratefromresearchlabstosoftwareengineeringcenters.Asagentsaredeployed
inapplicationssuchaselectroniccommerce,andstarttotakeoverresponsibilitiesfor
theirhumanusers,techniquesforcontrollingtheirautonomybecomecrucial.Aswell,
theavailabilityoftoolsthatfacilitatethedesignandimplementationofagentsystems
becomes an important factor in how rapidly the technology will achieve widespread
use.Consequently,theATAL2000programincludedtwospecialtracksonAutonomy—
Theory,Dimensions,andRegulation andAgentDevelopmentTools.Besidespapersin
eachofthesespecialtracks,theprogramalsofeaturedtwoassociatedpanels(organized
by Cristiano Castelfranchi and Keith Decker respectively). Another highlight of this
year’sprogramwastheinvitedtalksbyleadingexponentsofagentresearch:
architectures CraigBoutilier StructuredOnlineandOfflineSolutionof
Decision-TheoreticPlanningProblems
theories PhilCohen OnGroupCommunication
It is both our hope and our expectation that this volume will be as useful to the
agentresearchanddevelopmentcommunityasitssixpredecessorshaveprovedtobe.
WebelievethatATALandtheIntelligentAgentsseries,ofwhichtheseproceedingsform
apart,playacrucialroleinarapidlydevelopingfield,byfocusingspecificallyonthe
relationships between the theory and practice of agents. Only through understanding
theserelationshipscanagent-basedcomputingmatureandachieveitswidelypredicted
potential.
March2001 CristianoCastelfranchi
YvesLespe´rance
WorkshopOrganization
OrganizingCommittee
CristianoCastelfranchi UniversityofSiena,Italy
YvesLespe´rance YorkUniversity,Canada
SteeringCommittee
NicholasR.Jennings UniversityofSouthampton,UK
YvesLespe´rance YorkUniversity,Canada
Jo¨rgP.Mu¨ller SiemensResearch,Germany
MunindarP.Singh NorthCarolinaStateUniversity,USA
MichaelWooldridge UniversityofLiverpool,UK
ProgramCommittee
ChittaBaral (USA) SuzanneBarber (USA)
MichaelBeetz (Germany) StefanBussmann (Germany)
LawrenceCavedon (Australia) PaoloCiancarini (Italy)
PhilCohen (USA) RosariaConte (Italy)
GiuseppeDeGiacomo(Italy) KeithDecker (USA)
FrankDignum (TheNetherlands) AlexisDrogoul (France)
JacquesFerber (France) KlausFischer (Germany)
MichaelFisher (UK) StanFranklin (USA)
FaustoGiunchiglia (Italy) PiotrGmytrasiewicz(USA)
KeithGolden (USA) BarbaraGrosz (USA)
HenryHexmoor (USA) WiebevanderHoek (TheNetherlands)
MarcHuber (USA) Markd’Inverno (UK)
NickJennings (UK) DavidKinny (Australia)
SaritKraus (Israel) MichaelLuck (UK)
John-JulesMeyer (TheNetherlands) Jo¨rgMu¨ller (Germany)
AnandRao (UK) MurrayShanahan (UK)
OnnShehory (Israel) CarlesSierra (Spain)
MunindarSingh (USA) LizSonenberg (Australia)
KatiaSycara (USA) MilindTambe (USA)
JanTreur (TheNetherlands) ManuelaVeloso (USA)
TomWagner (USA) WayneWobcke (UK)
MikeWooldridge (UK) EricYu (Canada)
AdditionalReviewers
JaelsonCastro RinoFalcone AnujGoel
ClaudiaV.Goldman JohnGraham DanielGross
PuHuang LukeHunsberger ManuelKolp
DungLam FosterMcGeary WiekedeVries
WeiYang
TableofContents
SectionI:AgentTheoriesI
OptimisticandDisjunctiveAgentDesignProblems ........................... 1
MichaelWooldridgeandPaulE.Dunne
UpdatingMentalStatesfromCommunication .............................. 15
A.F.Dragoni,P.Giorgini,andL.Serafini
SensingActions,Time,andConcurrencyintheSituationCalculus ........... 31
StephenZimmerbaumandRichardScherl
SectionII:AgentDevelopmentToolsandPlatforms
DevelopingMultiagentSystemswithagentTool .............................. 46
ScottA.DeLoachandMarkWood
LayeredDisclosure:RevealingAgents’Internals ............................ 61
PatrickRiley,PeterStone,andManuelaVeloso
ArchitecturesandIdioms:MakingProgressinAgentDesign ................. 73
JoannaBrysonandLynnAndreaStein
DevelopingMulti-agentSystemswithJADE ................................ 89
FabioBellifemine,AgostinoPoggi,andGiovanniRimassa
SectionIII:AgentTheoriesII
High-LevelRobotControlthroughLogic .................................. 104
MurrayShanahanandMarkWitkowski
DeterminingtheEnvelopeofEmergentAgentBehaviourviaArchitectural
Transformation ........................................................... 122
OswaldoTera´n,BruceEdmonds,andSteveWallis
SectionIV:ModelsofAgentCommunicationandCoordination
DelegationandResponsibility ............................................. 136
TimothyJ.NormanandChrisReed
AgentTheoryforTeamFormationbyDialogue ............................ 150
FrankDignum,BarbaraDunin-Ke¸plicz,andRinekeVerbrugge
TaskCoordinationParadigmsforInformationAgents ...................... 167
MarianNodine,DamithChandrasekara,andAmyUnruh
X TableofContents
SectionV:AutonomyandModelsofAgentCoordination
PlanAnalysisforAutonomousSociologicalAgents ......................... 182
MichaelLuckandMarkd’Inverno
MultiagentBiddingMechanismsforRobotQualitativeNavigation .......... 198
CarlesSierra,RamonLo´pezdeMa`ntaras,andD´ıdacBusquets
PerformanceofCoordinatingConcurrentHierarchicalPlanningAgents
UsingSummaryInformation .............................................. 213
BradleyJ.ClementandEdmundH.Durfee
SectionVI:AgentLanguages
AgentProgrammingwithDeclarativeGoals ............................... 228
KoenV.Hindriks,FrankS.deBoer,WiebevanderHoek,
andJohn-JulesCh.Meyer
ModelingMultiagentSystemswithCASL–AFeatureInteraction
ResolutionApplication .................................................... 244
StevenShapiroandYvesLespe´rance
GeneralisedObject-OrientedConceptsforInter-agentCommunication ..... 260
RogierM.vanEijk,FrankS.deBoer,WiebevanderHoek,
andJohn-JulesCh.Meyer
SpecificationofHeterogeneousAgentArchitectures ........................ 275
SimoneMarini,MaurizioMartelli,VivianaMascardi,andFlorianoZini
SectionVII:Planning,DecisionMaking,andLearning
ImprovingChoiceMechanismswithintheBVGArchitecture ............... 290
LuisAntunes,Joa˜oFaria,andHelderCoelho
Planning-TaskTransformationsforSoftDeadlines ......................... 305
SvenKoenig
AnArchitecturalFrameworkforIntegratedMultiagentPlanning,
Reacting,andLearning ................................................... 320
GerhardWeiß
SectionVIII:PanelSummary:AgentDevelopmentTools
PanelSummary:AgentDevelopmentTools ................................ 331
JoannaBryson,KeithDecker,ScottA.DeLoach,MichaelHuhns,and
MichaelWooldridge
TableofContents XI
SectionIX:PanelSummary:Autonomy—Theory,Dimensions,andRegulation
AgainonAgents’Autonomy:AHomagetoAlanTuring
—PanelChair’sStatement ............................................... 339
CristianoCastelfranchi
AutonomyasDecision-MakingControl .................................... 343
K.SuzanneBarberandCherylE.Martin
Autonomy:Theory,Dimensions,andRegulation ........................... 346
RinoFalcone
SituatedAutonomy ....................................................... 349
HenryHexmoor
Autonomy:ANiceIdeainTheory ......................................... 351
MichaelLuckandMarkd’Inverno
AdjustableAutonomy:AResponse ........................................ 354
MilindTambe,DavidPynadath,andPaulScerri
AuthorIndex ............................................................. 357
Introduction
Like its predecessors [5,6,3,4,2,1], this volume of the IntelligentAgents series focu-
sesontherelationshipsbetweenthetheoryandthepracticeofintelligentautonomous
agents. The volume is divided into eight sections. Sections I and III present work on
agenttheories.SectionIIdiscussesworkondevelopmenttoolsandplatformsforagents
andmultiagentsystems.SectionIVpresentsworkonmodelsofagentcommunication
and coordination. SectionV deals with the issue of autonomy and presents additional
workonagentcoordination.SectionVIdealswithprogrammingandspecificationlan-
guagesforagents.SectionVIIfocusesonissuesinagentarchitectures,suchashowto
supportplanning,decisionmaking,andlearning.Finally,SectionsVIIIandIXprovide
summariesofthediscussionsheldatthetwoATALpanels,oneonAgentDevelopment
ToolsandtheotheronAutonomy—Theory,Dimensions,andRegulation.
SectionI:AgentTheoriesI
WooldridgeandDunnedevelopaformalmodeloftheproblemofdesigninganagentthat
achievesand/ormaintainsgivengoalsandprovesomenewresultsaboutthecomplexity
oftheagentdesignproblemundervariousassumptions.Theylookatoptimisticagent
design,wheretheagentisonlyrequiredtoachieve/maintainitsgoalsforsomeexecution
of the specified environment.They also look at disjunctive design, where the agent is
allowedtoachieve/maintainonlyoneamongasetofalternativegoals.
Dragonietal.developaformalframeworkformodelinghowagents’mentalstatesare
updatedasaconsequenceofcommunicationacts.Mentalstatesaremodeledasmulti-
contextpropositionaltheories.Anabductiveaccountofmentalstatechangeisdeveloped
fortheframework.Theuseoftheframeworkisillustratedbyformalizingthe“inform”
speechactandshowinghowitseffectsonthehearer’smentalstatecanbecaptured.
Zimmerbaum and Scherl propose a generalized version of a framework for reasoning
about how knowledge is affected by action, to deal with concurrent actions and time.
Theframework,whichisbasedonthesituationcalculus,capturestheknowledgeeffects
of both sensing and non-sensing actions. Indexical knowledge of time and absolute
knowledgeoftimearebothhandled.
SectionII:AgentDevelopmentToolsandPlatforms
DeLoachandWoodpresenttheirMaSEmethodologyformultiagentsystemdevelopment
and a tool that supports its use. The methodology is comprehensive, covering both
the analysis and design phases. Diagram-based notations are provided for developing
detailedmodelsofthebehaviorsassociatedwithrolesandconversations.
Rileyetal.presenttheirlayereddisclosureapproachtosupportingthedebuggingandfine
tuningofcomplexagentsystems.Inlayereddisclosure,thesystemdesignerdefinesan
informationhierarchy,whichisthenusedtocontrolthelevelofdetailoftheinformation