Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Agent, Services and Organization Oriented Analysis and Design ---- Building Open Enterprise Infrastructure Supporting Trading and Mining Longbing Cao Faculty of Information Technology University of Technology, Sydney, Australia Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 2 What’s the problem? • How does the problem emerge? – Project from Capital market CRC – Industrial requirements for capital markets and financial services – Bridge linking both industrial and research requirements – Problems having industry value and research value Industrial requirements-driven & interestingness-driven research 2017/5/22 [email protected] 3 What’s the problem? Datamining Program Team Leader: Prof Chengqi Zhang FIT, UTS: Broadway F-Trade Infrastructure (integrating DataSources, trading/mining algorithms, offers personalized services in terms of system, data and algorithms) Infrastructure Longbing & Jiarui’s work ITR&D-Enabled Finance Multi-agent & Data-mining Internet Multiple * Remote Data Sources CMCRC: CBD of Sydney (Industrial requirements; Users: Anybody, anytime, anywhere, from KDD & Finance; Services: System, algorithms, data) Industry Brokers, retailors Applications like Wanli’s work Researchers (Data mining, financial researchers, financial analysis, decision support analysis…) Australian Technology Park: Redfern; FIT,UTS (Diff. Providers: AC3, HK market, CSFB, etc. Diff. Formats: FAV, ODBC, JDBC, OLEDB, etc.) Data& resources 2017/5/22 Data mining Jiaqi & Li’s work [email protected] 4 What’s the problem? • What are specific problems from financial markets? – Evaluating trading and mining strategies from industry & research – Accessing real huge capital data crossing markets – Stock & rules association, selection, and optimization and integration – Pattern discovery in stock markets – Cross market analysis – Applications as investment decision support ITR&D-Enabled Finance & Teamwork 2017/5/22 [email protected] 5 What’s the problem? • What’s my specific problem? – Key linkage: build a comprehensive and powerful infrastructure • • • • • • Trading and mining supports Online, flexible, automated, enterprise-oriented, open Plug and play soft components Personalized and customized in different granularities Reporting & visualization looks like a virtual service provider – Automated Enterprise Infrastructure Supporting Trading and Mining in Capital Markets – Testbed of both research and applications for the project 2017/5/22 [email protected] 6 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 7 Objectives • Can trading and mining be supported in one system? – Differences between trading and mining – Mutual features or requirements for trading and mining support systems • Algorithms • Requirements for dataset, data pre-processing & postprocessing • Human system interaction could be similar • System and knowledge management could be unified • Software complexities are similar 2017/5/22 [email protected] 8 Objectives • What is expected for a system supporting both trading and mining? – first satisfy the above mutual features – support integration of heterogeneous and distributed data sources, and transparent to operational systems – support multiform of algorithms both from trading and mining, multiform of data sources, multiform of user types and profiles – Algorithms, system modules, user information, information and knowledge resource can be plugged into and removed from the system locally and remotely – automatically registration of algorithms or other components into the system – Variant user profiles and financial domain concepts can be supported – expandable for future finance-oriented research and applications – privacy of plugged algorithms can be kept 2017/5/22 [email protected] 9 Objectives Trading Services Mining Services Back-testing Data preprocessing Signal alert Feature selection Market replay Methods selection (like classification, regression, clustering, or others) Basic charting Training/testing/deploying process Draw tools Evaluation & refinement Reporting Knowledge presentation or visualization Technical analysis/ Fundamental Analysis Interpreting mined patterns Stock recommendation* Prediction or description Integration of trading strategies* Deployment in the real world* E-training/learning & applications* Method optimization* Automated execution* Integration of multiple methods* Cross markets* Multiple data sources* 2017/5/22 [email protected] 10 Objectives • What are main research objectives? – concrete functions can be available in the automated enterprise infrastructure – research methodology and methods are required for building such an infrastructure – what are key research problems, and what research values are there? How to solve these problems? – Supporting both research and development in academic and industrial projects in CMCRC project – Research papers and PhD thesis can be generated Agent service-oriented analysis and design 2017/5/22 [email protected] 11 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 12 Related work • System classification – Black Box, White Box, Glass Box, Grey Box • Similar systems – Computerized trading systems • TradeStation, E-Trade, TradeTech – Data mining • IM, EM, Clementine, Angoss, WEKA None of them can do the work we proposed. 2017/5/22 [email protected] 13 Related work • Research methods – Objects, components, services and agents – Object-oriented, component-based, service-oriented, agent-oriented 2017/5/22 [email protected] 14 Related work Object-oriented Component-based Agent-oriented Origin Semantic net Semantic net Symbolic AI & behaviour-based AI Computational entity Objects Components Agents State parameters for an entity unlimited unlimited Intentional stance, like belief, desire of an agent Activity Passive Passive Proactive with self control thread, automated Computational process Message passing and method response Message passing and method response Message passing and method response Message types Unlimited, handle messages by method vocation Unlimited, handle messages by method vocation Speech act; Language abstraction elements Objects, classes, modules Reuse, design patterns, application framework Agents, class, modules, design patterns, framework, organization, roles, society, goal Modelling abstraction mechanism Fine object as an action entity, method invocation used for describing interaction. static organizational modelling, no semantics More strong abstraction mechanism, e.g. component, reuse, design patterns, application frameworks, Agents as coarse and automated computing entity, social ability (organization, roles, etc), dynamic organizational modeling Analysis and design abstraction in fine granularity; Object model, dynamic model, function model; Abstraction in more coarser granularity; component library, framework library, object bus More coarser abstraction; Role model, interaction model, agent model, service model, acquaintance model encapsulation State and behaviour State and behaviour, application framework State and behaviour, behaviour activation Organizational relationship Static syntactic inheritance Static syntactic and structural inheritance An interactive network with inter- and intra-subsystem and subsystem elements interaction, multiple organizational relationships (hierarchy, marketing, etc) Interaction Syntactic interaction, invoking methods or functions, simple message passing Syntactic interaction, invoking methods or functions, message passing Interaction on knowledge and social levels System problem solving Event/behaviour-driven; design-time decision; no automated problem solving; predefined execution Event-driven; design-time decision making Goal-driven, automated and flexible problem solving; active decision making at runtime, reasoning ability Complexity in problem solving Generic system, with predefined interactive relationships Not strong enough for modelling complex systems Building complex distributed systems(data, ability, and control) System property Somehow encapsulation, autonomy, passivity and interaction Somehow encapsulation, autonomy, passivity and interaction Autonomy, reactivity, sociality, proactiveness; loose control, bigger freedom, uncertainty and indeterminism Evolution (specialization) of OO Evolution (specialization) of OO and CB 2017/5/22 Inter-relationship [email protected] 15 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 16 Research methodology • Research methodology in my work – Agent Service-Oriented Approach • Agent-oriented methodology + service-oriented architecture – agent service-oriented analysis and design 2017/5/22 [email protected] 17 Research methodology • Agent-oriented methodology – MASE Methodology – MESSAGE methodology – TROPOS methodology – Gaia methodology 2017/5/22 [email protected] 18 Research methodology Gaia Metaphor Organization Modeling late requirements engineering, modeling of the environment Analysis phase MASE MESSAGE TROPOS Organization Organization use-cases UML, AUML early and late requirements engineering role model, protocols application goals and subgoals, agent roles, interaction Structure, role, topological relations, Role, organizational structure, functional and nonfunctional requirements, structural dependencies Design phase global organizational rules agent classes, agent interaction protocols, system architecture control entity, workflow structure Agent systems Open agent systems closed agent systems Open agent systems Limitations early requirements analysis Open modeling the organizational rules, design the organizational structure 2017/5/22 [email protected] global laws 19 Research methodology • Agent services-oriented approach – Organization-oriented metaphor: abstraction – FIPA Abstract Architecture: architectural elements and their relationships – Organization-oriented modeling: RA – Agent-oriented methodology: analysis & design – Service-Oriented Architecture: architecture – Java Web Services: architecture & implementation – Java Agent Services: architecture & implementation agent service-oriented analysis and design 2017/5/22 [email protected] 20 Research methodology • Why agent services-oriented approach – large scale open agent-based system • • • • • 2017/5/22 Open Large scale Interoperable, enterprise applications-oriented web-based environment Service of quality: interactions, flexibility, autonomy, reliability, security [email protected] 21 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 22 Key research work • System functional & nonfunctional requirements • Goal-oriented organizational modeling • System architecture • Agent service ontology and semantic relationships • Agent service-oriented analysis & design 2017/5/22 [email protected] 23 Key research work • Representation and registry of agent services • Agent service directory • Agent service communication • Agent service transport • Mediation of agent services • Discovery of agent services 2017/5/22 [email protected] 24 Key research work • System functional & nonfunctional requirements – Data services support – Algorithm services support – System services support – Trading support – Mining support – Quality of service, development objectives or architectural constraints 2017/5/22 [email protected] 25 Key research work • Goal-oriented organizational modeling – Visual modeling • extended i* framework – Formal modeling • first-order logic + scenario analysis – Integrative modeling • Visual modeling + Formal modeling 2017/5/22 [email protected] 26 Key research work Integrative Modeling: Hybrid representation with visual modeling and formal specifications Visual Modeling: Informal box-and-arrow notations, minimal syntax, an ontology, no semantics Formal Modeling: Formal conceptual model, with assertion language for specifying and constraints Informal box-and-arrow notations, minimal syntax, no ontology, no semantics Formal box-and-arrow notations, with an ontology, syntax, and semantics 2017/5/22 Integrative Modeling Formal conceptual model, with an assertion language for specifying rules and constraints Global qualities like flexibility, availability, security, adaptability, etc. [email protected] Formal Modeling Informal Modeling Nonfunctional Requirements Functional Requirements Informal box-andarrow notations, minimal syntax, with or no ontology, with or no semantics Organizational model includes actors, goals and inter-dependencies 27 Key research work 2017/5/22 [email protected] 28 Key research work 2017/5/22 [email protected] 29 Key research work • Service-oriented architecture – organizational framework & design patterns Service Requesters 2017/5/22 ServiceNet Bind h lis Pub Fin Mat d/ ch Service Brokers (Dispatchers) Service Providers [email protected] 30 Key research work • EAI architecture – Administration Center, Algorithms Center, Control Center, Services Center, and User Center F-TRADE User Roles F-TRADE Applications Subscriber Users Technical Analyses University Coordinators CMCRC Users Public Users Administrator Users Algorithms Providers Fundamental Analyses Risk Analyses Cross Market Mining Investment Analyses Stock Data Services F-TRADE Function Centers Administration Center Algorithms Center Control Center Services Center User Center Web Server F-TRADE Services Support Control Engine W e b S e r v e r Organization Framework Design Patterns Data Services Algorithm Services F-TRADE Gateway Agents Data Resource Interface& Operation Gateways /Adapters Algorithm Mediator Agent Metadata Management Data Mining Algorithm Base Trading Signals Algorithm Base Ontology Base System Services Algorithm Interface Agent F-TRADE Data &System Resources System Resource Interface& Operation Gateways /Adapters Data Mining & Trading Algorithms in Stock Markets Knowledge Base Local DataSource Legacy Systems Users DataSource AC3 DataSource WSDL/UDDI/SOAP &Internet & Intranet & Extranet Infrastructure 2017/5/22 [email protected] 31 Key research work • Agent service ontology and semantic relationships – Ontology profiles • Domain ontology • Task-method ontology • Ontological commitment – Ontological engineering • Agent service ontology, Ontology specifications, semantic relationship, Ontology transformation, Naming of agent services, Representation of agent services 2017/5/22 [email protected] 32 Key research work Financial Organization Bank Savings And Loans ... Exchange Credit Union Exchange Type Price Futures Index Option Stock OTC Exchange Vendor Exchange Exchange Exchange StockMarket Auction Market ... OpenPrice ClosePrice BidPrice AskPrice TradePrice Dealer Market Fixed Income Stock su bc pa ta la rtnc ss of e-o of f in s 2017/5/22 Money Market Futures FinancialOrder Limit Order LimitPrice ... Instrument Type Foreign Index Exchange ... Market Order Options Price Enter Dealer FinancialOrder ... Stop Order MarketOrder LimitOrder OrderOperation Amend StopPrice StrikePrice StockPrice Trade Delete Date Time StopOrder Cantr Volume [email protected] 33 Key research work ;; definition of LimitOrder (subclass LimitOrder FinancialOrder) (documentation LimitOrder "An order to a &%Broker to buy a specified quantity of a &%Security at or below a specified price, or to sell it at or above a specified &%limitPrice.") ;; definition of bidPrice (instance bidPrice TernaryPredicate) (domain bidPrice 1 Object) (domain bidPrice 2 CurrencyMeasure) (domain bidPrice 3 Agent) (documentation bidPrice "(bidPrice ?Obj ?Money ?Agent) means that ?Agent offers to buy ?Obj for the amount of ?Money.") (=> (bidPrice ?Obj ?Money ?Agent) (exists (?Offering) (and (instance ?Offering Offering) (patient ?Offering (exists (?Buying) (and (instance ?Buying Buying) (agent ?Buying ?Agent) (patient ?Buying ?Obj) (transactionAmount ?Buying ?Money))))))) 2017/5/22 [email protected] 34 Key research work • Agent service-oriented analysis & design – Role Model, Interaction Model, Environment Model, Organizational Rules, Organizational Structure, Agent Model, Service Model, and Agent Service-oriented Architecture 2017/5/22 [email protected] 35 Key research work Goal RegisterAlgo InformalDef When an algorithm component has been coded and the algorithm isn’t available from the system at the moment, this algorithm component can be registered into the system by calling plug-in interfaces, filling in algorithm registration ontologies, and upload the algorithm module. FormalDef Actor Provider Mode achieve Attribute constant ca: CodeAlgo Attribute constant algo: Algorithm registered: boolean Creation condition ● Fulfilled(ca) ¬ Existed(algo) Invariant ca.actor = actor Fulfillment condition ac: AlgorithmComponent (ac.algo = algo t1 cpi: CallPluginInterfaces (cpi.actor = actor Fulfilled(cpi) pi.Called) t2( faro: FillinAlgoRegisterOntologies (faro.depender = actor Fulfilled(faro) aro.Filled) 2017/5/22 [email protected] 36 Key research work CodeAlgo (ca1) Created Fulfilled AlgorithmComponent (ac1) CallPluginInterfaces (api1) Created Fulfilled Created FillinAlgoRegisterOntologies (faro1) Fulfilled Created UploadAlgoComponent (uac1) Fulfilled Created True False AlgorithmComponent (ac1).registered t0 2017/5/22 [email protected] t1 t2 t3 t4 t5 37 Key research work Role Schema: PLUGINPERSON Description: This preliminary role involves applying registering a nonexistent algorithm, typing in attribute items of the algorithm, and submitting plug in request to F-TRADE. Protocols and Activities: ReadAlgorithm, ApplyRegisteration, FillinAttributeItems, SubmitAlgoPluginRequest Permissions: reads Algorithms // an algorithm will be registered changes AlgoApplicationForms // algorithm registration application form changes AttributeItems // all attribute items of an algorithm Responsibilities Liveness: PLUGINPERSON = (ReadAlgorithm).(ApplyRegisteration). (FillinAttributeItems)+.(SubmitAlgoPluginRequest) Safety: The algorithm agent has been programmed by implementing AlgoInterface agent and ResourceInterface agent, and is available for plug in. This algorithm hasn’t been plugged into the algorithm base. 2017/5/22 [email protected] 38 Key research work • Representation and registry of agent services – Namespace and service root – specifications for agents and services registration – representation and registration management • micro-level + macro-level 2017/5/22 [email protected] 39 Key research work AgentService RegisterAlgorithm(algoname;inputlist;inputconstraint;outputlist;outputconstraint;) Description: This agent service involves accepting registration application submitted by role PluginPerson, checking validity of attribute items, creating name and directory of the algorithm, and generating universal agent identifier and unique algorithm id. Role: PluginPerson Pre-conditions: -A request of registering an algorithm has been activated by protocol SubmitAlgoPluginRequest -A knowledge base storing rules for agent and service naming and directory Type: algorithm.[datamining/tradingsignal] Location: algo.[algorithmname] Inputs: inputlist InputConstraints: inputconstraint[;] Outputs: outputlist OutputConstraints: outputconstraint[;] Activities: Register the algorithm Permissions: -Read supplied knowledge base storing algorithm agent ontologies -Read supplied algorithm base storing algorithm information Post-conditions: -Generate unique agent identifier, naming, and locator for the algorithm agent -Generate unique algorithm id Exceptions: -Cannot find target algorithm -There are invalid format existing in the input attributes 2017/5/22 [email protected] 40 Key research work • Agent service directory – agent directory service • agents directory entries • Discovery of agent directory-entries – service directory service • service directory entries • Discovery of service directory-entries – specification of directory service – position of agent service directory 2017/5/22 [email protected] 41 Key research work public interface AgentDirectory extends Directory{ AgentDescription[] getAgentDescription(); Vector getDirectoryEntry(); void register(AgentDescription ad) throws DirectoryFailure; void update(AgentDescription ad) throws DirectoryFailure; void delete(AgentDescription ad) throws DirectoryFailure; void execute(AgentDescription ad) throws DirectoryFailure; AgentDescription[] search(AgentDescription ad) throws DirectoryFailure; void setDirectoryEntry(Vector de); } 2017/5/22 [email protected] 42 Key research work • Agent service communication – communication model, communicative act, and communication control in agent and service communication – Message-based communication model, agent service message model 2017/5/22 [email protected] 43 Key research work • Agent service transport – representation and transport of messages – agent service message model – transport protocol – specifications for transport service 2017/5/22 [email protected] 44 Key research work public interface AgentTransport extends Transport{ AgentLocator getSender(); AgentLocator getReceiver(); String getTransportType(); Locator getTransportAddress(); Message getTransportMessage(); void setSender(AgentLocator sender); void setReceiver(AgentLocator receiver); void setTransportType(String ttype); void setTransportAddress(Locator taddress); void setTransportMessage(Message tmessage); } 2017/5/22 [email protected] 45 Key research work • Mediation of agent services – mediation and management of agent services • local meditation • global mediation • multi-tier mediation strategies – mediation protocols – mediation logic 2017/5/22 [email protected] 46 Key research work • Framework & patterns 2017/5/22 [email protected] 47 Key research work • Discovery of agent services – search for an ontology – query an agent or service • agent/service directory service – search for a message • search for original message or encoded message 2017/5/22 [email protected] 48 Key research work Component OntologySearchService Description: This component deals with the search of a corresponding ontology concept in target ontology space according to a user defined key word or ontology term from source concept space in user profile. Pre-conditions: -Ontology spaces enclosing the possible target ontology concepts must be prepared -A knowledge base storing all existing matching rules -A knowledge base storing transformation rules Services: -S1: UserProfileTransformer //other properties are omitted for limited space -S2: UserProfileMatcher //other properties are omitted for limited space - S3: AutomaticOntologyMatcher -Actor: OntologySearchService -Role: Based on ontology concepts in ontology space of user profile, look for corresponding ontology concepts in target ontology space -Pre-conditions: -Ontology concepts in ontology space of user profile found by service UserProfileMatcher -Activity: search -Permissions: -Read supplied knowledge base storing transformation rules -Read supplied knowledge base storing matching rules from ontology concepts in user profile to concepts in target ontology space -Post-conditions: -Find existing matching records, or -Find and output ontology concepts of target ontology space -Store new query matching rule into knowledge base if available -Similar value: simValue = 1 -Exception: Cannot find relevant ontology concepts in target ontology space -S4: ManualOntologyMatcher //other properties are omitted for limited space Post-conditions: -Output ontology concepts into target ontology space Exceptions: -No existing target ontology space -No knowledge base storing transformation rules -No knowledge base storing matching rules 2017/5/22 [email protected] 49 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 50 Significance and contributions • Research value – investigation of agent services-oriented analysis and design for building open enterprise infrastructure – some interesting research work • agent services-based modelling, integrative modeling, domain-specific ontologies, ontological engineering • representation, directory, transport, transformation, mediation and discovery of agent services – research testbed and platform • multi-agent, high frequency data mining, cross market mining, stock stream data mining 2017/5/22 [email protected] 51 Significance and contributions • Application value – Industrial requirements from CMCRC – A prototype supporting trading and mining • Development from researches – financial, data mining, intelligent systems, multi-agent… • Benefit users – brokers, retailers, data applicants, services applicants… • Applications testbed and platform – data mining, stock stream data processing, technical analysis, fundamental analysis, risk analysis, investment risk, market replay, signal alerts, cross market mining… 2017/5/22 [email protected] 52 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 53 Evaluation • Functional & nonfunctional evaluation – As an agent-based system • Distributed and Interoperable, Flexible, Open and Dynamic, Automated, User-friendly, Adaptive, Privacy-keeping – As a trading/mining support system • Supporting plug in and online registration of data sources, system modules, and algorithms • Providing data gateway for supporting data linkage and cross markets • Supporting user profiles-oriented and problem domainoriented interaction • Supporting online system customization and reconstruction • Supporting comprehensive add-on applications from capital markets 2017/5/22 [email protected] 54 Evaluation • Empirical evaluation – system comparison – practical testing in the real world – computational performance – customer feedbacks 2017/5/22 [email protected] 55 Function Items TradeStation F-TRADE Trading supports Back-testing Yes Yes Indicators Yes Yes, but not enough Basic charting Yes Yes, but not strong Draw tools Yes Can support Reporting Yes Yes, but not strong Signal alert Yes Can support Automated execution Yes No Market replay No Can support Stock recommender No Can support E-training & learning No Can support Technical analysis Yes Yes Fundamental Analysis No Can support Data formats Some commercial stock data JDBC, ODBC, FAV Intraday Yes Yes Inter-day Yes Yes Real time Yes No Downloadable No Yes Cross markets N/A Yes Web-based Yes Yes Downloadable pack Yes N/A Formula disclosed No Can support Rule customizing Yes, using EasyLanguage N/A Rule optimization Yes Yes Rule integration No Yes Inner language Yes, EasyLanguage N/A Data plug in No Yes Module plug in No Yes Algorithm plug in No Yes Interface creating Yes Yes, automated Data gateway No Yes, data link and access System reconstructing No Yes Multi-level outputs N/A Yes Visual output Yes Yes, not strong Data supports System supports 2017/5/22 System logging Yes Privacy keeping N/A [email protected] Yes, Algorithm, system Yes 56 Content • • • • • • • • What’s the problem? Objectives Related work Research methodology Key research work Significance and contributions Evaluation Conclusions & Future work 2017/5/22 [email protected] 57 Conclusions & Future work • An automated enterprise infrastructure integrating both stock trading and data mining • Agent services-oriented approach as design paradigm for building open agentbased systems • Agent services-driven enterprise infrastructure supporting trading and mining 2017/5/22 [email protected] 58 Conclusions & Future work • Teamwork • Refinement of agent services-oriented analysis and design • System implementation • Publications • Thesis preparation 2017/5/22 [email protected] 59 List of Publications • • • • • • • • • • Longbing Cao, Jiarui Ni, Jiaqi Wang, Chengqi Zhang. Agent Services-Driven Plug-in Support in F-TRADE. 17th Australian Joint Conference on Artificial Intelligence, December 2004, Queensland, Australia. Longbing Cao, Dan Luo, Chao Luo, Li Liu. Ontology Discovery in Multiple Ontology Domains. 17th Australian Joint Conference on Artificial Intelligence, December 2004, Queensland, Australia. Longbing Cao, Chao Luo, Dan Luo, and Chengqi Zhang. Integration of Business Intelligence Based on ThreeLevel Ontology Services. proceedings of the 2004 IEEE/WIC/ACM International Conference on Web Intelligence(WI'04),IEEE Computer Society Press, 20-24 September, 2004, Beijing, China. Longbing Cao, Jiaqi Wang, Li Lin, and chengqi zhang. Agent Services-Based Infrastructure for Online Assessment of Trading Strategies. proceedings of the 2004 IEEE/WIC/ACM International Conference on Intelligent Agent Technology (IAT'04), IEEE Computer Society Press, 20-24 September, 2004, Beijing, China. Longbing Cao, Chao Luo, Dan Luo, Li Liu. Ontology Services-Based Information Integration in Mining Telecom Business Intelligence. Proceeding of PRICAI04, Springer Press, 2004. Longbing CAO, Dan LUO, Chao LUO, and Chengqi ZHANG, Systematic Engineering in Designing Architecture of Telecommunications Business Intelligence System, Design and Application of Hybrid Intelligent System (Proceedings of International Conferencec on Hybrid Intelligent System), Melbourne, Australia, 14-17 Dec 2003. pp. 1084-1093. (ISBN: 1 58603 3948 [IOS Press]). Longbing Cao, Chao Luo, Chunsheng Li, Chengqi Zhang, and Ruwei Dai, Open giant intelligent information systems and its agent-oriented abstraction mechamism, In: Proceedings of the fifteenth International Conference on Software Engineering and Knoledge Engineering (SEKE 2003), San Francisco, California, USA, July 1-3, 2003. pp.85-89. (ISBN: 1-891706-12-8) Li Lin, Longbing Cao, Jiaqi Wang, Chengqi Zhang, The Applications of Genetic Algorithms in Stock Market Data Mining Optimisation, Proceedings of Fifth International Conference on Data Mining, Text Mining and their Business Applications, September 15-17, 2004, Malaga, Spain. Longbing Cao, Chunsheng Li, Chengqi Zhang, and Ruwei Dai, Open Giant Intelligent Information Systems and Its Agent-Oriented Analysis and Design, Proceedings of The 2003 International Conference on Software Engineering Research and Practice (SERP'03), Vol.2, pp. 816-822, Las Vegas, Nevada, USA, June 23-26, 2003. CSREA Press. (ISBN: 1-932415-20-3). L.B. Cao, R.W. Dai. Agent-Oriented Metasynthetic Engineering for Decision Making, International Journal of Information Technology and Decision Making, 2(2):197-215, World Scientific Publishing, 2003. 2017/5/22 [email protected] 60 Thank you for your attention! Comments & suggestions? 2017/5/22 [email protected] 61