ѧ
JOURNAL OF SOFTWARE
1999ꡡ109ڡVol.10No.91999



һֻڰԶAgentЭͬƲβģ
ܡڿ
ժҪηӳƹЭͬǿЭͬϵͳٵһҪ.һЭͬƵĲβģͣagentĳjoint endeavor̬ԪԼһagent֮ͨԭ淶ģжagentΪ.ֻjoint endeavorЭͬģʽӳЭͬƵһЩ.»ֲβģ͵һʵֽṹ򵥽һЭͬͼϵͳCODDS(collaborative object diagram design system).
ؼʡЭͬЭͬƣ֪agentintentionjoint endeavor.
ͼšTP391
A Hierarchical Operation Model Based on Half-automated Agent for Collaborative Design
LI Bao-jian,LIN Zong-kai
(CAD Laboratory Institute of Computing Technology The Chinese Academy of Sciences Beijing 100080)
ZENG Guang-zhou
(Department of Computer Shandong University of Technology Ji'nan 250061)
AbstractHow to embody the requirements of collaboration in designing processes is an important issue in developing collaborative design systems. In this paper, a hierarchical operation model for collaborative design is presented. A mental component called joint endeavor and a set of communication primitives are devised for the half-automated agents in the model to meet some of the requirements in collaborative design. An implementation structure for the hierarchical operation model is also given. A collaborative object diagram design system (CODDS) is introduced as an example of this collaborative design model.
Key wordsCSCW, computer supported collaborative design, cognitive agent, intention, joint endeavor.
ڼЭͬƣcomputer supported collaborative design,CSCDĸʽУЭرҪ,ڴĿиˣΪĿ漰ͬ,Ҫͬר֪ʶͼ.ΪCSCWһرʽCSCDעԱЭͬڷָǹļ1,2.ִĲƷоЭͬƵý弼֧֣Զָ롢װ塢Ӧùߵȣ3,4.agentӳЭͬԼ,Ӷõ֧֡ԼЭͬƹе-˽оټ.M.R.Patraˣ5ݶЭͬagent̬ȣjoint attitudesıҪԼ佨ϵĳ̽.D.J.Ramˣ6һֻԼԪconstraintmeta-objectģͣԼԪֲͬԱƶ֮ĲԼӶԲֵطӳЭͬƵΪЭͬṩһ֧.
һЭͬƵĲβģ.ֲβģͨԱϵͳ֮ðԶagentЭͬƵԼ.ǽͨЭͬƹеһЩ淶agentΪ,ӶģͻЩһagent̬ԼһӦagentͨԭӦagentʵֽṹ.ΪֲβЭͬģ͵һʵĻǿһЭͬͼϵͳ.
1ЭͬƵĻ
һЧCSCDϵͳ뿼ЭͬƵԼ56ݣǿ԰
ֽ⡡ЭͬƲߵĶѣԽֽɲͬȵ.ַֽĿڽÿһӦר.ȻҲһЩҪͬļҪ.ֽδһԵɣǿܷ֣һֽɸС.
ɫ䡡ƳԱƹаݲͬĽɫЩɫĻܵ.費ͬ߻СȻвͬĽɫͬĶ߿ܾ֯ߡٲߡעߵȲͬĹܽɫȻҲܾȫԵȵĹϵ.ԱСĽɫһɲģƹ̵ĽչͬһȫܰݲͬĽɫ.һЭͬϵͳӦṩƶȡȨĽɫģ.
ƻءһ񱻷ֽɲͬ񲢱費ͬߣǵƹ̱ܵأֱ.ƹовͬĽɫǵΪǵĽɫΥɫΪõⲢȡԱкӰ.
ͻ⡡ƹпܻͻ.вṹill-structured,SimonΪڽṹãwell-structured7.ͬһԱܶνֽвͬǾĳһƣҲвͬķ.ϵͳӦṩͻЭ̻.
װ䡡һɣӦڱһԵĻװ.װĹУܻ.һˣһ򼸸Ľ뱻޸ģ߱ǣҲ·ֽǵĸ.
ЭͬƵЩԼνǼɵһϵͳ.
2ЭͬƵĲβģ
ЭͬƵĲβģͷֳ3.ϲǶԱ²ǹƶмɶԶagentɵĶagentϵͳ.3֮Ĳϵ.ÿ߿ֱӲagentֻͨмagentӵزݹƶ.
2.1Agent
AgentΪǵ֪Ԫأcognitive component̬Ԫأmental componentbeliefs,desiresgoals,intentionscommitmentsȾ.̬ԪصѡȡShohamΪûһΨһȷѡҲûһѡġȷۣΪͬӦÿҪ̬ͬʣ8.Cohen Levesqueintentionjoint intention ۣ9ԼJenningsǵĹϵjoint responsibilityģͣ10ݷӳ˺һЩʣǵ۱Ե󾫲ӦƵӦñ.ˣjoint endeavor ĸһagentͨԭ淶ЭagentΪʹЭͬĳЩԼ.
2.1.1Joint endeavor
Joint endeavorģ̬ʱ̬߼ʽض.ǣоĿָЭͬϵͳԣǽʽ.Joint endeavoragentһָ̬Ԫأ䶨򵥵̬Ԫأbelief,mutual belief,achievement goal,joint achievement goal,persistent goal,joint persistent goal,intentionԼjoint intentionȣ910.
ĻϣǿԶjoint endeavorĸ.һagent,AG=(a1,a2,...,an),agentjoint endeavor pΪ(J-ENDEAVOR AG p)ҽ
(1) pΪһjoint achievement goal.
(2) ΪʵĿһܲԣͨЭЭ̽ĿpֽĿ(p1p2,...,pm)1mn,pi(i=1,...,m)һagent AGi=(ai1,ai2,...,aiMi)M1+M2+...+Mm=n.
(3) m=1ûзֽ⣬ôAGеagentΪʵʹpЭֱͬ
(3.1) ǹͬpѾΪ.ңpǸǱ뽫˽֪ͨpֵagentС飻pǸǵjoint endeavorɹ.
(3.2) ǹͬŲʹpΪ.pǸǱ뽫˽֪ͨpֵagentС飻pǸǵjoint endeavorʧ.
(3.3) Ǳֵ֪ͨڣʧ.
(4) m>1(ֽǷƽ)ôi1 m(J-ENDEAVOR AGi pi)AGi жpiֱӸǵĶвȨ.ÿjoint endeavorһֱеóɹʧܵĽֱĳjoint endeavorʧܣȻAGipiϵĹͣ.
κjoint endeavorѾʧܣpagentСAG ܾp·ֽǿΪpΪ.ڵ2£ΪpΪ棩pǸǱ뽫˽֪ͨpֵagentС飻pǸǵjoint endeavorʧ.
еjoint endeavorǳɹģAG˶ɹĹͬôǾͽװ׶.װʧܣAGܾp·ֽ⣬ǿΪpΪ棬˽֪ͨpֵagentС飨ڵĻװɹôĿpѾʵ֣AGӦ˽֪ͨpֵagentС飨ڵĻ.pǸô͵õjoint endeavor ս.
Joint persistent goal涨һagentӦʵĳĿĳŵֱΪĿѾʵֻ򲻿ʵ֣ȻJoint persistent goalûй涨ʵĿ꣨Ʃ˵ǽȴȥʵֻԼĳЩʵ֣.CohenLevesquejoint intentionҪagent ӦԼĳЩʵǵĿ.joint endeavorȻһjoint persistent goal涨ЩagentԼĳЩʵĿ꣬һ涨ʵĿһֲι滮¹滮.
2.1.2 Agentͨԭ
Agent֮ͨԭʵֽ.ǵĵӦñƵͨԭinformagree/denypermitforbidrequestconsent/refuse,proposal accept/decline. 
ͨԭָ˼Ϊ.̬ԪصѡΨһģΪͨԭѡҲΨһ.ǵĵӦñagentͨԭЩͨԭǶԷagentЧӦ.
Inform,agree/deny
(INFORM x y p)ʾagent x ֪ͨyp.agent xţbelievepԽ֪ͨһagent yҪagent뽲滰⽫±֪ͨagent yšagent xp.Ҳagent yp뷢Ϣ(AGREE y x p)x⽫xšypagent yp뷢Ϣ(DENY y x p)x⽫xšyp.
ͨԭ԰һagentͬ(mutual belief).
Request,consent/refuse
Request󣩷ӳagentdesire.һagentagentĳЩaction.actionָ򼶲Ĳβģйв.(REQ x y act-expr)ʾagent xagent yact-exprжʽй涨action.жʽactionaction.򼶲˵actionҲĶ.
agent yյϢ(REQ x y act-expr)Ϣ(CONSENT y x act-expr)(REFUSE y x act-expr)Ӧ.ǰߵyact-expr涨actionintention(CohenLevesequeINTEND19ݵʹ)Լagent xyintentionߵyľܾԼxyľܾ.
Permit,forbid
ǵĲβģУһagentһagentһôǾжԴֱӸǵĶвȨ.˴һʼagentСرֹЩκβ.ȻagentagentСagentagentСľͼܶǵǰ߿permit߶ǵĶĳЩ.
(PERMIT x y act-expr)ʾagent xagent yact-exprй涨Ĳ.Ȼact-exprй涨ĶxΧ֮ڣЧϢypermission.
(FORBID x y act-expr)ʾagent xֹagent yact-exprй涨.Ȼact-exprй涨ĶxΧ֮ڣֹֽЧϢyforbid.
Proposal,accept/decline
һagentֽ⡢װһֽܼʱǿܾвͬ.ھ֮ǰһҹͬܵķһϺõİ취.
һagentԶС齨ĳ.(PROPOSE x p)ʾagent xС齨鷽p.Ϣ͵гԱ.һԱyյ飬Ϣ(ACCEPT y x p)Ϣ(DECLINE y x p).xԱյacceptϢ֪ͨгԱϳŵѾγ.
2.1.3ЭͬĳЩ֧
мĶagentͨγһЭԵ̬֮ͨԭӶʹǵΪԹ淶к.мĶagent붥ĶԱһһӦ.ԱԵײϵͳĲԼ֮ĽΪͨмagentУagentļͰӶʹԱΪͬagent̬ͨԭ涨.
Աjoint endeavorΪֱ֧Эֽͬװ󣬶̬ȻҪԱĽɫǵƻмأǲܽΥǽɫĲ.̬ҪĳŵĺԽм.磬ǻ֪ǵһֵѾʧܣҪͣǵĹֱõǸ·ֽĽΪֹȻ·ֽĽжǵǷѾı䣬ѾĹǷҪ޸.
һԱֽ⡢װ߶ĳһֽܼмʱǿͬķͻ.ݴˣвͬĳͻⷽ.Э̾һЭͬгͻʵΪСвͬԱвͬרžͼǿĳЩǲʵ.
мagentЭͬĳЩԼ󣬶ܴԱоƾ.ˣagent֮ԶЭ̹Ա֮ĳͻ.ԱԽagent֮ͨԭproposal-accept/decline֮Э̹̽й淶.
˵һյһߣͨǵagentһproposalʱ߱acceptdeclineֻѡ񣬲ұѡ.ͨŹһֱеһͬķΪֹ.˿˵Ա֮Э̹һֻproposal-accept/declineͨЭЭ̷.˵Э̹һΪ̣ĴԼ.
2.2Ա
ǵģУκļϵͳΪЩĲ.仰˵ƵĲƷɶһ˵,С󹹳ɵģвͬĲ紴޸ļɾȵȿʩЩ.ˣƶֳΪϵͳdomain level system.һ˵ƻджɱеԱǹһͨǲĶʩӲģõյԱ.
ԱɹһܵԱ.ʱͬ磬Ƕĳһзֽʱ.ʱҲ첽ع磬ڲͬԱһûбҪͬʱ.мagentЭͬĳЩ󣬲ҶԱγĳԼмagentѡͽֹԱĳЩ飩ⲢζԱҪԼ.мagentԽֹһԱĳֲǿĳֲȻѻԱȥ.ζԱĺ̬ȶƻĳɹǷǳҪ.
3ϵͳʵо
ǸβģеİԶagentͼ1ʾϵṹ.ͼ1УΪ˷㣨ϵͳҲagentһȻagentһ.

ͼ1Զagentϵṹ
ǵ漰ӦãԶagentû.Agentû汻ֳ֣ûԪû.û֧Ա͹ĽԪû֧Աͨagentĺ.
ģְǸԱͨûԹĲ.Աjoint endeavorĽɫǰƽ׶Լĵǰ״̬жԱĿܲǷʵ.ȷĲݵϵͳΥĲ򵥵رûֹ.
ЭͨģԪûֱϵ֧ͨԪû˽.ЭͨģĹܱȼģĹܸӵö.Эͨģ3Ҫ.
1 joint endeavor.仰˵ܸjoint endeavorγɡԼֹ.ǰjoint endeavorķʽ߼УΪͻص㣬ǽagentĸ޶Ϊnƹnǲģ.ھʵֹУǵʵԣµagent˵Ա뵽ĳС飬ĳСĲԱ˵agent˳.joint endeavorûʵӰ.
2 ʵֵ2.1.2agent֮ͨԭͨЭ.ΪЩͨԭԱͨagentԪûֱʹõģЭѡʹͨԪûҪĺ.磬һյĳֲһʱԻش.شЭ齫ֹκ.ˣЭԹ淶ԱĽΪ.
3 joint endeavorԼЭͬ.ƵĲͬ׶ΣЭͬjoint endeavorĹ涨һ.ЭģͬǿԱѭjoint endeavorԼ.磬һǸڵԱƹ󣬸joint endeavor뽫ǵĹֵ֪ͨԱ.agentЭģͨдݽṹԵضȡı仯ʵһ.һagentЭģͨȡ⵽һֵʧܣӦͣص丸.
AgentҲάŵǰagentڲʾ.ֱʾҲǷǳҪģΪһagent֪ǰ״̬agentǰ̬ȷѡ;.
AgentԹĲͨģʵֵ.ͨʽʽʽȶַʽ.ģĹܼʵַŷʽĲͬŲ죬ĹҪagentԹĲ.
4һЭͬͼϵͳ
Ȼjoint endeavorΪĵЭͬƲβģһͨõģͣĿǰΪֹǽһЭͬͼϵͳCODDSЭͬģ͵Ч.
ķOOAҪĽ׶.OOAмսöͼʾעϸʹöڱĲβģеĹĸ.ںߣǽǿݵķװԣ.ͼԡṹװṹӦģ.ͼֱӷӳ󣬲Ҳƺʵֽ׶ʹöϵת.
OOAһӵĹ.ͬأֱ۷ӳOOA̵ĶͼƹҲһӵĹ.ΪҪ棬ϵͳԱרԼûȵ֪ʶ.κγɹOO˵ЭǱزٵ.CODDSĹУȫ˱ЭͬƲβģͼʵּ.
CODDSϵͳۿһͨͼα༭༭ĻԪǴԺͶιϵװϵľΡıߵ.ΪͼڲıʾΪ洢ʽ.ŵڣС˴洢ұڽͼڲ֮ͬĹ.ͼÿһĸϢ񡢴Լ߶Ȩ޵ȵȣԷز뵽ͼıʾ֮.мagentֱʹЩϢЭͬ.
5
һֻڰԶagentЭͬƲβģ.Ϊģ͵ĺģ˶agentһָ̬Ԫءjoint endeavorԼһͨԭ淶agentΪӶЭͬƵĳЩ.joint endeavorЭͬƣǹ涨ԱһЧЭͬģʽҲְԶagentֲβģ͵һʵֽṹ.
ΪЭͬƲβģ͵һӣǿһЭͬͼϵͳCODDS.CODDSĳʹþģǿеġЧģдøľģͣЧ.
ЭͬƲβģһ־һͨԵģ.ģӦǱǾ޴.ǽרҺƳЭͬϵͳԳַģ͵Ǳ.
*оõɽʡȻѧ.
ͨѶϵˣ250061,ɽҵѧϵ
߼飺1964ʿҪоΪ˹ܣЭͬ.
ܣ1947ڣҪоΪ˹֪ܺʶ̣ѧ.
ڿ1934оԱʿʦҪоΪݿ⼰Эͬ.
ߵλ ڿйѧԺ㼼оCADоʵ  100080
ܣɽҵѧϵ  250061
οף
1Ellis C, Gibbs S, Rein G. Groupware some issues and experiences. Communications of the ACM, 1991,34(1):3858
2Grudin J. CSCW introduction. Communications of the ACM, 1991,34(12):3034
3Fluckiger F. Understanding Networked Multimedia Applications and Technology. Englewood Cliffs, NJ: Prentice Hall, Inc., 1996
4Gibbs S, Verrijn-Stuart A A. Multi-User Interfaces and Applications. Amsterdam: North-Holland Publishing Company, 1990
5Patra M R, Mohanty H. Building joint attitudes for collaborative design in multi-agent environments. In: Siriruchatapong P, Lin Zong-kai, Barthes J P eds. Proceedings of the 2nd International Workshop on CSCW in Design. Beijing: International Academic Publishers, 1997. 209214
6Ram D J, Vivekananda N, Rao C S et al. Constraint meta-object: a new object model for distributed collaborative designing. IEEE Transactions on System, Man and Cybernetics, 1997,27(2):208220
7Simon H A. The structure of ill-structured problems. Artificial Intelligence, 1973,(4):181201
8Shoham Y. Agent oriented programming. Artificial Intelligence, 1993,(60):5192
9Cohen P R, Levesque H J. Intention Is Choice with Commitment. Artificial Intelligence, 1990,(42):213261
10Jennings N R. Controlling cooperative problem solving in industrial multi-agent systems using joint intentions. Artificial Intelligence, 1995,(75):195240
ոڣ1998-07-13޸ڣ1998-09-21
