Dynamic Agent Communities Facilitating to Distant Learning in a Virtual University Information Space Vadim Ermolayev Dept. of Mathematical Modelling and Information Technologies, Zaporozhye State University, 66, Zhukovskogo st., 330600, Zaporozhye, Ukraine, tel/fax:+380 61 264 17 24, E-mail: eva@zsu.zaporizhzhe.ua Motivation and Background VUIS It’s like CA Real World or Microsoft Encarta I guess I know how to use this!!! It’s like doing Hierarchy of distributed Layered mediator my everyday work, wrapped legacy IS University IS but electronically!!! with unified interfaces and information resources VUIS is considered to be a passive media inhabited by active “beings” (agents) representing real life actors… EVA: IS2000-VUDE, 06.11.2000 2 1
Active VUIS Inhabitants Persistent MAS and Member Agents University MAS R [ \ U Proxy Agent(s) S VU MAS Cloning Agent SDS Coordination R [ \ S U Agent V Dept. MAS Middle Middle Agent Agent Ontology Agent Generic MAS agents Lower level V Dept. MAS Proxies “wrap” respective MAS and are the representative members in the higher level MAS EVA: IS2000-VUDE, 06.11.2000 3 Agents’ Interactions – Parametric Feedbacks a = f ( X , Y ) a – accept a new student A j A i to the class ~ ~ ~ ~ = Y Y ( y ( X ),..., y ( X )) ~ − Professor’s interest 1 n y 1 X ( ) in new students 1.0 Parametric feedbacks – ~ − Entry test 0.8 expressed attitude , y n ( X ) difficulty level capability to perform 0.6 requested action 0.4 (policy) at a certain state 0.2 5 10 15 No of students at the class EVA: IS2000-VUDE, 06.11.2000 4 2
Agent Communities as dynamic coalitions for performing tasks (sets of works) Why? – there are MAS already - tasks emerge and are soon accomplished - agents join the coalition if they are capable to contribute to the task execution - an agent may participate in more than one coalitions at a time How? – accept or reject the work, react with parametric feedbacks What is changing in time? - the coalition – agents join and move out - agent’s state – capability to perform a work - agent’s knowledge (beliefs) about the capabilities of the other MAS members EVA: IS2000-VUDE, 06.11.2000 5 Framework Models Communication Model Generic Agent Model Functional System / Component Model Coordination Model Process Model ???? Evolution Model ???? EVA: IS2000-VUDE, 06.11.2000 6 3
Agent’s Evolution Model Agent’s proactive self-development and self-adaptation in response to the changes in MAS and environment caused by task(s) execution Capabilities Beliefs S i w A 1 ~ j ← Y f ( s , X ) j i c 1 ... S j w ~ A i B c j ( X ) A w Y i j ... ∈ s , s S j c i j A n i = A n s { r ( X ), q ( F ), t ( F )} A A [ ] r ( X ) - parameter constraints 0 , 1 A q ( F ) - policy constraints A t ( F ) - transition function EVA: IS2000-VUDE, 06.11.2000 7 Case Study – VUDE Domain PhD recruiting scenario: Agent Phase 1: A PhD candidate submits the CV Proxy Proxy Agent and indicates his/her intention to become a PhD (secretary) Tutor student ( PA , CA , TA ) Tutor Agent Agent (User) (User) Cloning Phase 2: The CV is analysed and the best Agent Professor Match is searched ( TA , PRA -s) Phase 3: Qualified candidate passes the test SDS Coordination from the chosen professor ( TA , PRA , PA ) Agent Phase 4: Successful candidate is interviewed Middle Middle Agent Agent and assigned to a research project ( TA , PRA ) Ontology Agent Phase 5: The professor and his assistant Department MAS External influences and prepare the individual curriculum for the Co-ordination MAS reactions accepted candidate as well as the list of Facilitator cloning Ontology queries Influences Task communities recommended reading ( TA , PRA , AA , CMA , Professor ( PRA ), Middle Assistant ( AA ), Course master ( CMA ), LA ) Agent Librarian ( LA ) agents EVA: IS2000-VUDE, 06.11.2000 8 4
VDept Agents and their Roles (w.r.t. the Case Scenario) Proxy Agent: PA – accept outer influences, order to clone TA , pipeline the candidate to his TA Utility Agents: CA – Clone a TA each time Z aporozhye State U niversity a new PhD candidate comes Inst. of M athem atics and Cybern etics P R 2 P R 3 P R 4 Dept. O f M athem atical M odelling and IT S ecretary M rs Proxy P E lena Vetukh to PA Faculty: C hair Prof. Vyachyslav Tolok PR 1 TA PA Prof. PR 2 COA – Monitor agents activities, Sergey B orue Prof. Vadim E rmo layev PR 3 PRA5 A ss. P rof Sergey G om enyuk A 1 Accept work results to its SDS, A ss. Prof Vitaly M ukhin A 2 A ss. Prof Nadezhda M a tv eyshina A 3 A ss. Prof Nataly a K eberle A 4 Provide work params on request. A ss. Prof Vladim ir D avidovsky A 5 A ss. Prof M axim Shum akov A 6 Virtual F aculty OA – Provide common dept. Visiting: Prof. PR 4 Victor G ristcha k A 4 CM A ontologies on request AA1 PRA1 PRA1 A 5 Virtual: LA Prof. Anatoly D oroshenko P R 5 A 2 A 3 L A 6 Middle Agents: PRA1-PRA5 – evaluate candidates’ qualifications; provide a test, evaluate test results, arrange the interview, provide course recommendations AA1 - AA6 – prepare working plan, prepare curriculum CMA – provide electronic courses, issue calls for new courses LA – provide electronic textbooks, issue calls for new teaching materials EVA: IS2000-VUDE, 06.11.2000 9 Modelling ( How did they play their roles? ): Phase 3. Testing, Step 1 – Test is ordered. t : n execute : t w ∅ No Results 1 TA: Require_Test policy: w = Requre_Tes t X Y W ( , , } w − w 1 1 1 1. Generate 4 works TA 2 5 2. Invoke TA to influence ∅ No works for redirection TA w 2 , w PRA i with 4 w 3 , w TA itself with ~ = 5 d g W W W { , } TA TA TA PRA i w = (' Provide _ Test ' , X , Y ), 2 2 2 w = Test X Y (' ' , , ), 3 3 3 w = (' Evaluate _ Re sults ' , X , Y ), 4 4 4 w = TA (' Analyse _ Marks ' , X , Y )} 5 5 5 Just monitoring… CoA COA EVA: IS2000-VUDE, 06.11.2000 10 5
Modelling ( How did they play their roles? ): Phase 3. Testing, Step 2 – Test is provided. t : t : n + n 1 t execute : delay w 3 , w w 5 1 w redirect to itself 3 TA TA W w = TA (' Test ' , X , Y ), w 3 3 3 w = 5 (' Analyse _ Marks ' , X , Y )} 5 5 5 execute, delay TA g W w w TA 2 4 redirect to itself PRA i PRA i w w = (' Evaluate _ Re sults ' , X , Y ), 2 4 4 4 w 4 PRA i : Provide_Test policy: ~ ~ 2 2 = 1. Generate test file ( ) Y Results: Y { Test _ Form PRA i PRA i =< > according to the ontology FILENAME } and received params X 2 Monitoring 2. Invoke PRA i to influence CoA and accepting feedback(s)… COA with: ~ COA 2 ( ‘Accept_Results’ , ) Y PRA i EVA: IS2000-VUDE, 06.11.2000 11 Modelling ( How did they play their roles? ): Phase 3. Testing, Step 3 – Test is performed. t : t : : t + n + n 1 n 2 t execute, delay execute : delay w w w 3 , w w 3 5 5 1 w redirect to itself 3 TA TA TA W w = TA (' Analyse _ Marks ' , w 5 5 X , Y )} w 5 5 execute, delay delay TA 4 w w 2 4 redirect to itself PRA i w PRA i PRA i w = 4 (' Evaluate _ Re sults ' , 4 X , Y ) TA: Perform_Test policy: 4 4 ~ 2 ~ 3 = = 1. Require and accept = Results: Y TA { Test _ Results Params: X Y { Test _ Form 3 PRA i w =< > parameters from COA =< > FILENAME } FILENAME } 3 2. Invoke human user to Monitoring, fill in the test form. accepting feedback(s), CoA 3. Invoke TA to influence providing work parameters… COA COA with: ~ 3 ( ‘Accept_Results’ , ) Y TA EVA: IS2000-VUDE, 06.11.2000 12 6
Modelling ( How did they play their roles? ): Phase 3. Testing, Step 4 – Test results are evaluated. t : t : t : t : + + n + n 1 n 2 n 3 t delay delay execute,delay execute : w w 3 , w w w w 5 5 3 5 1 redirect to itself w TA TA TA TA w = W 5 (' Analyse TA 5 _ Marks ' , w execute execute,delay delay X , Y )} TA 4 w w w 5 5 4 2 4 w PRA i PRA i PRA i 4 ~ 3 = = Params: X Y { Test _ Results 4 TA =< > FILENAME } ~ i = 4 Results: Y {( m ,... m ), PRA 1 k Monitoring, ( s ,... s )} 1 k CoA accepting feedback(s), providing work parameters… COA EVA: IS2000-VUDE, 06.11.2000 13 Modelling ( How did they play their roles? ): Phase 3. Testing, Step 5 – Parametric Marks are analysed. t : t : t : t : t : + + + + n n 1 n 2 n 3 n 4 t delay execute,delay delay execute execute : w w 3 , w w w w w 5 5 3 5 5 1 Negative: inform w TA TA TA TA TA W 5 w = TA (' Inform g W 7 TA on _ failure ' , execute execute,delay delay w TA ∅ 4 w w w X , Y )} 4 2 4 7 7 No Results PRA i PRA i PRA i Positive: Interview – Phase 4. = w (' Require _ Interview ' , 6 = TA X { X , X }, Y )} 6 1 5 6 ~ = i = 4 Params: X Y {( m ,... m ), 5 PRA 1 k ( s ,... s )} 1 k Monitoring, CoA providing work parameters… COA EVA: IS2000-VUDE, 06.11.2000 14 7
Recommend
More recommend
