2008 IEEE Congress on Services Part II
A Framework for Distributed Market Place based on Intelligent Software Agents and Semantic Web Services Ali Khalili, Ali Habibi Badrabadi, Farid Khoshalhan IT Engineering Group, K.N.Toosi University of Technology, Tehran, Iran
[email protected],
[email protected],
[email protected] heterogeneity related problems in a distributed business environment. Therefore, by integrating intelligent software agents and semantic web services, we can create a dynamic and autonomous environment for handling e-commerce interactions in a distributed market place. In this paper, we aim to create a framework for automation of a distributed marketplace on the web through agents. The proposed framework includes a semantic web service that advises managers for making strategic decisions and managing business models. We believe that a combined use of our proposed framework together with creativity and expertise of managers can improve decision making in knowledge intensive markets. The rest of the paper is organized as following. In section 2, we provide a general overview of Agent Technology. In section 3, we provide an overview of the Semantic Web Services and in particular WSMO and its reference implementation namely WSMX and finally in section 4, we describe our proposed framework for distributed market place and introduce StraBM semantic web service architecture.
Abstract In this paper we present a framework for automation of distributed market place that is based on intelligent software agents and semantic web services. A multi-agent environment designed for automation of e-commerce activities such as negotiation between buyer and seller, recommendation, and customer decision making. Furthermore we use a semantic web service middleware called StraBM , advising managers of e-shops for making appropriate strategic decisions and managing business models. StraBM uses a customized WSMX middleware to generate appropriate strategy and business models.
1. Introduction E-commerce involves complex processes with many facets, spanning areas that cover business modeling, information technology and social and legal aspects [18]. The bloom of e-commerce has changed the whole outlook of traditional trading behavior. Through the Internet, different business entities, including suppliers, retailers, and consumers, can now easily interact with each other and have their transactions within a minimum time [19]. Two general purpose of e-commerce is interoperation and automation, whereby in most cases there is a dependency of automation upon interoperation. Systems developed to cater the ecommerce process are using different platforms, standards, representations, and languages. However, there is an increasing need for these systems to interoperate; communicating information from one system to another; to have a better and efficient business environment. Furthermore, Heterogeneity poses great difficulties in realizing interoperation. In order to resolve this issue, software agents are deployed to automate many steps in the e-commerce process, which in turn can minimize cost and consequently maximizing profit [20]. Semantic web services are another solution for dealing with
978-0-7695-3313-1/08 $25.00 © 2008 IEEE DOI 10.1109/SERVICES-2.2008.11
2. Agent Technology Since the field of agent-based modeling is fairly new, no general agreement on the term agent has yet been established. A general definition of the term agent is a self-contained, problem-solving entity [14]. Also a software agent is defined as a software entity with its own thread of control able to execute operations without being externally invoked [14]. While in other place a software agent is defined as an agent that interacts with a software environment by issuing commands and interpreting the environment’s feedback. A software agent’s effectors are commands meant to change the external environment’s state. A software agent's sensors are commands meant to provide information [15]. Perhaps the most general way in which the term agent is used is to denote a hardware or (more usually)
141
are the most notable attempts to design a generalpurpose agent language. There exist several types of negotiation strategies; for example some kinds of negotiation mechanisms are based on Case-Based Reasoning (CBR) approach [25], fixed price [24][18], English auction [18][25][21], Dutch auction [18][21], iterative bargaining (Bilateral Contracts) [18][21][19] and Genetic Agent (GA) negotiation [24].
software-based computer system that enjoys the following properties: Autonomy: agents operate without the direct intervention of humans or others, and have some kind of control over their actions and internal state [15][16][13]; Social ability: agents interact with other agents (and possibly humans) via some kind of agentcommunication language [16][15][13]; Reactivity: agents perceive their environment, (which may be the physical world, a user via a graphical user interface, a collection of other agents, the Internet, or perhaps all of these combined), and respond in a timely fashion to changes that occur in it [16][15][13]; Pro-activeness: agents do not simply act in response to their environment; they are able to exhibit goal-directed behavior by taking the initiative [16][15][13]; Mobility: is the ability of an agent to move around an electronic network [16][15]; Veracity: is the assumption that an agent will not knowingly communicate false information [15]; Benevolence: is the assumption that agents do not have conflicting goals, and that every agent will therefore always try to do what it has been asked [15]; Rationality: is the assumption that an agent will act in order to achieve its goals, and will not act in such a way as to prevent its goals being achieved [15].
2.2. Recommendation While the rapid development of digital techniques has sped up the creation and invention of consumer electronic products, more and more advanced products now come to the market in relatively short period of time. Also there is a trend to integrate multiple functions into a single composite product, for example, the new type of mobile phone that can picture and transmit digital image in addition to transmitting voice. In this situation, when a consumer wishes to purchase a product, he has to spend more and more time to survey and compare relevant product information in order to find the product that best fulfills his personal needs. A promising solution to the above problem is to develop recommendation systems able to interact with consumers to capture their needs and to help them determine what to buy [19]. There are different types of product selection and recommendation approaches. Some important approaches are: Automated Collaborative Filtering (ACF) [25][20], Knowledge-Based (KB) [25], and hybrid [25]. ACF approach is based on “word-of-mouth” recommendations. This addresses the problem of capturing previous customers’ recommendations/feedback on the products they have already purchased, and use this feedback in recommending the products to potential new customers. KB approach is based on the knowledge-base of product information. Most of the knowledge-based systems for product selection and recommendation are either Case-Based Reasoning (CBR) [25][19] or GoalBased Retrieval (GBR) [25][19] type. CBR is a problem solving approach based on past experience. Past experience is organized in the form of cases, and is used to solve new problems. GBR approach is used to find products similar to ones the user is already aware of. The basic idea is “similarities are goalbased”. For example, an umbrella is similar to a raincoat if the goal is to protect from rain. But it is similar to a stick when the goal is to protect from a dog! For each goal, a similarity metric is defined
2.1. Negotiation Negotiation is a process of reaching an agreement on the terms (such as price and quantity) of a transaction for two or more parties. The negotiation process typically goes through a number of iterations, and in each of which one of the parties proposes an offer and sees whether the other accept. If not, other parties can propose their counter-offers and the process repeats until a consensus is reached [24]. Negotiation is a process with the goal of intended benefit, in which the buyer and the seller bargain resources such as price, product features, etc [25]. Two main issues involved in the agent-based automated negotiation are the communication between buying and selling agents, and the negotiation strategies used by the agents. The former is to do with agent-based languages and protocols that enable the message interchange between different agents. Some work has been reported in the area of agent communication among which knowledge query and manipulation language (KQML)[19][15], knowledge interchange format (KIF)[15] and FIPA Agent Communication Language (ACL) [19][7][23]
142
the first time, s/he might not understand their effect well. Assigning weights in the presence of several dozens of attributes is a difficult task and also traditional forms of weight elicitation such as pricing out or swing weighting which are traditionally used with MAUT do not work well in these cases [22]. Another approach used in commercial bid analysis products is optimization such as integer programming, and/or constraint programming. Bid analysis products from Emptoris and Rapt belong to this category. These products recommend a set of bids from multiple suppliers that optimizes one or more objectives, e.g., minimizing the total cost, set by the user. A drawback of this approach is that its capability for recommending a combination of bids is limited by its inability to express multiple objectives for optimization, which is necessary for the procurement of complex products. While this approach can be effective for simple objectives such as minimizing the total cost, it does not work well if the objectives involve complicated business rules over multiple attributes. Most importantly, this approach is weak in explaining its analysis results. These optimization-based systems do not allow the users to easily inspect the results and navigate through the bid information space, understand the properties of different bids, and compare the given options [22].
between each pair of products and the metric gives how close the two products are, in terms of the goal. Hybrid approach is primarily a combination of ACF and KB approaches. In some systems, the ACF is used in the post processing stage and the systems are predominantly knowledge-based. Other systems check whether there is sufficient number of feedbacks from previous users. This number is used as a threshold to decide what approach to be used. If the number is less than the threshold then a KB approach is used, otherwise ACF approach is used. The threshold value can be determined interactively based on the product and the business. Some systems attempted to utilize CBR approach for ACF. Some other systems are based on discrimination nets of product features or quantitative decision support tools. Apart from ACF, KB, and hybrid approaches, there are other product selection and recommendation approaches such as interviewing the user in order to know pair-wise feature importance, forming user models based on observation of user decision making, either in response to system suggestions (“candidate/revision” or “learn-on-failure”) or through passive observation of the decisions, and learning customer preferences by observing customers’ selections from return sets [25].
2.3. Agent-based Decision Making
3. Semantic Web Service Technology
Bid analysis products from Frictionless Commerce and Perfect use traditional decision analysis techniques, which have been actively studied in multiattribute decision making (MADM) [22][19], an area of operations research. The primary techniques in the field are Multi-Attribute Utility Theory (MAUT) [22][19], Simple Multi-Attribute Rating Technique (SMART) [22] and the Analytic Hierarchy Process (AHP) [22], all of them implemented in several software applications. The essence of all these widely used decision aids is breaking complicated decisions down into small pieces that can be dealt with individually and then recombined in an additive manner. The key difference is the way the scores on individual attributes and their weights are assessed. Perfect and Frictionless Commerce are based on MAUT. They request a user to assign relative weights to individual attributes of alternatives (i.e., bids), and then use an additive value function in order to compute the scores of the alternatives. The system then rank the alternative bids by score and the user selects the winning bids among the top-rankers [22]. A fundamental weakness of these packages is that they provide little guidance for weight assessment. Consequently, the resulting bid scores may not be reliable. When a user assigns weights to attributes for
Web services are loosely coupled software components which are published, located and invoked across the web, and accessible via standard web protocols [1]. They offer an interoperability model that abstracts from the idiosyncrasies of specific implementations [2]. Web Services are based on the following industry standards: Simple Object Access Protocol (SOAP), Web Services Description Language (WSDL), and Universal Description, Discovery, and Integration (UDDI). SOAP is an XML based lightweight messaging protocol intended for exchanging structured information between applications in a decentralized, distributed environment. WSDL is the W3C recommended language for describing the service interface. As services become available, they may be registered with a UDDI registry, which represents a set of protocols, can subsequently be browsed and queried by other users, services and applications [3]. Today’s Web was designed primarily for human Interpretation and use. The semantic Web is an extension of the current Web in which information is given well-defined meaning, better enabling computers and people to work in cooperation [4]. The main
143
element of semantic web is Ontology. Ontology is defined as a “formal, explicit specification of a common conceptualization” [5]. In other words, Ontology is a conceptualization of an application domain in a human-understandable and machinereadable form, and typically comprises the classes of entities, relations between entities and the axioms which apply to the entities in that domain [3]. Adding semantic web annotations to web services will enable the automation of various kinds of tasks, including discovery, composition, and execution of web services in an open, unregulated, and often chaotic environment (that is, the Web)[1].There are several approaches to define semantic Web services, the most prominent proposals are OWL-S and WSMO. OWL-S specifies a set of ontologies for the description of web services. Web services ontologies based on OWL-S are composed of service profile, service model, service grounding [6]. The service profile tells "what the service does", in a way that is suitable for a service-seeking agent (or matchmaking agent acting on behalf of a serviceseeking agent) to determine whether the service meets its needs. The service model tells a client how to use the service, by detailing the semantic content of requests, the conditions under which particular outcomes will occur, and, where necessary, the step by step processes leading to those outcomes. A service grounding ("grounding" for short) specifies the details of how an agent can access a service. Typically grounding will specify a communication protocol, message formats, and other service-specific details such as port numbers used in contacting the service. Web Service Modeling Ontology (WSMO) defines the modeling elements for describing Semantic Web services based on the conceptual grounding set up in the Web Service Modeling Framework (WSMF) [7], wherein four main components are defined: ontologies, Web services, goals, and mediators. WSMO inherits these four top elements, further refining and extending them [8]: • Ontologies represent a key element in WSMO since they provide (domain specific) terminologies for describing the other elements. They serve a twofold purpose: defining the formal semantics of the information, and linking machine and human terminologies. • Web services describe the functional behavior of an actual Web services.
•
•
Goals specify objectives that a client might have when consulting a Web service, i.e. functionalities that a Web service should provide from the user perspective. Mediators describe elements that aim to overcome the mismatches that appear between the different components that build up a WSMO description.
In this paper, we choose to use WSMO as our semantic web service model because of its explicit support of mediators and its powerful execution environments.
3.1. WSMX Environment)
(Web
Service
Execution
WSMX is a reference implementation for WSMO, designed to allow dynamic discovery, invocation and composition of Web Services. WSMX offers complete support for interacting with Semantic Web Services. In addition, WSMX supports the interaction with nonWSMO, but classical Web Services ensuring that a seamless interaction with existing Web Services is possible [2].
4. The Proposed Framework Our e-commerce model extends and builds on the ecommerce structures presented in [18]. Basically, our environment acts as a distributed marketplace that hosts e-shops and allows e-clients to visit them and purchase products. Clients have the option to negotiate with the shops, to bid for products and to choose the shop from which to make a purchase. Conversely, shops may be approached ”instantly” by multiple clients and consequently, through negotiation mechanisms (including auctions), have an option to choose the buyer. Additionally, each of the e-shops' managers who has registered to this environment can use the StraBM semantic web service that helps him for making strategic decisions and managing business models. The top level conceptual architecture of the system illustrating proposed types of agents and their interactions in a particular configuration is shown in Figure 1. Four negotiation protocols have been applied: FIPA English auction, FIPA Dutch auction, iterative bargaining and fixed price (also known as take-it-or leave- it). Note that the first two are one-to-many negotiations while the last two are one-to-one negotiations.
144
set of feasible alternatives, which the decision maker wishes to maximize. This function aggregates the criteria x1,...,x n. Besides, individual (single-measure) utility functions U1(x1), ..., Un(xn) are assumed for the n different attributes. The utility function translates the value of an attribute into "utility units". The overall utility for an alternative is given by the sum of all weighted utilities of the attributes. For an outcome that has levels x1, ..., Xn on the n attributes, the overall utility for an alternative is given by The alternative with the largest overall utility is the most desirable under this rule. The assessment of weights is a core issue when MAUT is used. Often, some kind of subjective judgment forms the basis for the weights, and yet the interpretation of the weights is not always clear. Many different methods exist for assessing the attribute values and weights such as pricing out, swing-weighting approach and pairwise comparisons of all attributes in AHP approach. ABSolute provides WORA (Weight determination based on Ordinal Rankings of Alternatives), a new approach where decision makers only provide ordinal rankings over subsets of bids. The visual interface of ABSolute helps users examine and select subsets of offers, and create and arrange ordinal rankings over these subsets. From the information implied by these ordinal rankings, the system derives a set of weights of attributes, and then an overall ranking of all the given offers using optimization techniques. With additional information from the decision maker, these results are iteratively refined. Simulations show that after only a few iterations WORA generates very good estimates of the decision maker's true weights. Let us now describe each agent appearing in that figure and their respective functionalities. A Client agent (CA) is created by the Personal agent to act within the marketplace on behalf of a user that attempts to buy ”something.” Similarly, a Shop agent represents user who plans to sell ”something” within the e-marketplace. After being created both Shop and Client agents register with the CIC agent to be able to operate within the marketplace. Returning agents will receive their existing IDs. In this way we provide support for the future goal of agent behavior adaptability. Here, agents in the system are able to recognize status of their counterparts and differentiate their behavior depending if this is a”returning” or a”new” agent that they interact with. There is only one Client Information Center (CIC) agent in the system. It is responsible for storing, managing and providing information about all ”participants” existing in the system. To be able to participate in the marketplace all Shop and Client agents must register with the CIC agent, which stores information in the Client Information Database (CICDB). The CICDB combines
Figure 1. E-Commerce Platform Architecture A hybrid method for recommendation [19] has been applied, in this system, two categories of knowledge, including the explicit and the tacit, are used to build knowledge acquisition and behavior-matching abilities for seller agents. The explicit knowledge is collected from the domain experts, which is codified and embedded into the system to evaluate the quality of various kinds of products. This involves a process of deriving and capturing the opinions of product experts and then converting them into a specially designed internal form for knowledge representation. On the other hand, the tacit knowledge denotes the experiences of user–system interactions collected from the previous users. The consumer experiences here describe the process of how they found the ideal products under the guidance of the system. The ABSolute Approach [22] has been applied for client decision making, ABSolute provides an integrated approach to support complex purchasing decisions and an opportunity to satisfy analysis capabilities required in large-scale purchasing processes. This approach integrates methods from multi-attribute decision analysis with visualization techniques. The core components of the ABSolute decision analysis framework are: • A user interface for interactive visual analysis; • MAUT - a traditional and widely used decision aid and; • WORA - a new methodology designed to determine weights in the presence of a large number of criteria. The most widely used technique used to decide between alternatives with multiple objectives is the Multi-Attribute Utility Theory (MAUT). The basic hypothesis of MAUT is that in any decision problem, there exists a real valued function U defined along the
145
the function of client registry, by storing information about and unique IDs for all users and of yellow pages, by storing information about of all shops known in the marketplace. Thus Client agents (new and returning) communicate with the CIC agent to find out which stores are available in the system at any given time. A Client agent is created for each customer that is using the system. Each Client agent creates an appropriate number of ”slave” negotiation agents with the ”buyer role” (Buyer agents hereafter). One Buyer agent is created for each store, within the marketplace, selling sought goods. On the supply side, a single Shop agent is created for each merchant in the system and it is responsible for creating a slave negotiation agent with the ”seller role” (Seller agent hereafter) for each product sold by the merchant within her e-store. There are two databases in the system: a single CICDB database (operated by the CICDB agent containing the information about clients, shops and product catalogues, and a single Shop Database (ShopDB) operated by the ShopDB agent storing information about sales and available supplies for each merchant registered within the system. The central part of the system operation is comprised by price negotiations. Buyer agents negotiate price with Seller agents. For this purpose Buyer agents migrate to the e-stores known by the CIC agent to carry sought after commodity. In case of multiple Buyer agents attempting at purchasing the same item, they may compete in an auction. Shop agent send results of price negotiations to the Client agent that decides where to attempt at making a purchase. Another agent that you can see in this framework is Manager Agent. This agent is responsible for helping e-shop's manager strategic decision making and business model managing. Our proposed system to support the manager in the strategic decision making process is like a hybrid of an ES and a DSS [9][17]. In this scenario Manager Activates Manager Agent and makes choices on the strategic Decision-making task required. Manager Agent as a service requester connects to StraBM semantic web service through the WebAgent adapter that uses WSMX adapter to solve problems related to communicating with StraBM. After service match making and composition (if necessary) in the StraBM semantic web service, Manager Agent recommends the outcome to the manager and waits in the background for new decision of manager. If the manager does not accept the outcome, Manager Agent iteratively tries to satisfy the manager.
StraBM is our proposed Web Service for advising appropriate strategy and business models to managers and decision-makers of firms. It helps managers to make long-term decisions and choose right business models in different conditions of market. StraBM is based on a semantic web service model that Intelligent Software Agents can easily interact with it and employ its capabilities.
Figure 2. StraBM Web Service Environment StraBM uses a customized WSMX middleware to generate appropriate strategy and business models. The main components of StraBM as depicted in Figure 2 are: 1. Business Models Ontology (BMO): BMO is foundation for a variety of management tools that facilitate business decisions. It describes the architecture of the firm and its network of partners for creating, marketing and delivering value and relationship capital, in order to generate profitable and sustainable revenue streams. Use of e-business models is essential in an increasingly dynamic and uncertain business environment for the following reasons [11]: a. Helps identifying and understanding the relevant elements in a specific domain and the relationships between them. b. Enables knowledge representation and helps agents easily communicate and share their understanding of an e-business among other agents in the decision making process. c. Mapping and using e-business models as a foundation for discussion facilitates change. Business model designers can easily modify certain elements of an existing e-business model. d. A formalized e-business model can help identifying the relevant measures to follow in an e-business.
4.1. StraBM Semantic Web Service
146
In this paper, we use business models ontology that is introduced in [11]. The proposed BMO is influenced by the Balanced Scorecard approach and emphasizes on the following four areas that a business model has to address: a. PRODUCT: What business the company is in, the products and the value propositions offered to the market. b. CUSTOMER INTERFACE: Who the company's target customers are, how it delivers them products and services, and how it builds strong relationships with them. c. INFRASTRUCTURE MANAGEMENT: How the company efficiently performs infrastructural or logistical issues, with whom, and as what kind of network enterprise. d. FINANCIAL ASPECTS: What is the revenue model, the cost structure and the business model’s sustainability.
4.
5.
6. The Balanced Scorecard is a management concept developed in the early 90s that helps managers measure and monitor indicators other than purely financial ones. 2.
Strategic Decision Maker (SDM): The Strategic Decision Maker component interacts with decision-makers, receives choices and judgmental inputs from them and assigns strategy-making tasks to particular web services.
5. Conclusion
When activated, the marketing strategy web service performs reasoning and recommends marketing strategies based upon McDonald’s (1989b, 1996) fourbox directional policy matrix. The global marketing strategy web service produces strategic advice on global marketing strategies by using Harrell and Kiefer’s (1993) nine-cell model. The internet/ecommerce strategy web service generates strategies by applying Watson and Zinkhan’s (1997) four-cell matrix.The competitive strategy and associated IT/IS strategy web service searches for and makes suggestions on competitive strategies through using Porter’s (1980) generic strategy model and Pearlson and Saunders’s (2004) guidelines for IT/IS strategies [9]. After determining appropriate strategy, if decision maker agent requests for business models, SDM sends strategy related information to BMM for generation business models. 3.
Furthermore, BMM updates the BMO by interacting with market place. WSMX Resource Manager: The Resource Manager is responsible for management of repositories to store definitions of Web Services, goals, ontologies, mediators and judgmental inputs for decision making on strategies and business models. WSMX Manager and Execution Engine: Execution management has the role of coordinating the execution of the application services in a meaningful scenario. Such scenarios are called “Execution Semantics”, and they can be seen as low-level projections of the requirements of the application layer. The execution semantics define the way the application services have to work together to complete a useful scenario. Having the execution semantics as a separate concept in the architecture is the first step towards a highly decoupled system [10]. Adapter: Adapters address problems related to communicating with WSMX. They transform format of a received message or even extracted data from an API (Application Programming Interface) into the WSML compliant format understood by WSMX. WSML messages are encapsulated to WSDL and sent using SOAP protocol to WSMX. We call this transformation grounding.
This paper has introduced a framework for automation of distributed market place based on intelligent software agents and semantic web services. The proposed framework uses a multi-agent environment for automation of e-commerce activities such as negotiation between buyer and seller, recommendation, and customer decision making. Additionally a semantic web service middleware called StraBM has been used for advising managers of eshops to make appropriate strategic decisions and manage business models.
6. References [1] E. Paikari,J. Habibi, and S. H. Yeganeh, “Semantic Composability Measure for Semantic Web Services”, Proceedings of the First Asia International Conference on Modelling & Simulation (AMS'07) , Thailand, 2007, pp. 8893. [2] A. Haller,E. Cimpian, A. Mocan,E. Oren, and C. Bussler, “WSMX - a semantic service-oriented architecture”, Proceedings of the 2005 IEEE International Conference on Web Services, USA, 2005, pp. 321- 328 vol.1.
Business Models Manager (BMM): BMM performs reasoning using BMO and recommends Business Model(s) to decision-makers.
147
[3] J. Zhou, J. P. Koivisto, and E. Niemela, “A Survey on Semantic Web Services and a Case Study”, 10th International Conference on Computer Supported Cooperative Work in Design, IEEE, 2007,pp. 1-7. [4] T. Berners-Lee,J. Handler, and O. Lassila, “The Semantic Web”, Scientific American, 2001. [5] T. R. Gruber, “A Translation Approach to Portable Ontology Specifications”, Knowledge Acquisition, Academic Press, Stanford University, 1993, pp. 199-220. [6] W3C Web-Ontology Working Group, “OWL-S: Semantic Markup for Web Services”, http://www.daml.org/services/owl-s/1.1/overview/. [7] D. Fensel, and C. Bussler, “The Web Service Modeling Framework WSMF”, Electronic Commerce Research and Applications, Vol. 1, Elsevier B.V, 2002, pp. 113-137. [8] C. Feier, D. Roman, A. Polleres, J. Domingue, M. Stollberg, and D. Fensel, “Towards Intelligent Web Services: Web Service Modeling Ontology (WSMO) ”. In Proc. of the Int'al Conf on Intelligent Computing (ICIC) 2005, Hefei, China, August 23-26, 2005. [9] S. Li, “AgentStra: an Internet-based multi-agent intelligent system for strategic decision-making”, Expert Systems with Applications, Elsevier Ltd., Vol. 3, 2007, pp. 565-571. [10] D. Fensel, H. Lausen, J. de Bruijn, M.Stollberg, D. Roman, A. Polleres ,and J. Domingue, Enabling Semantic Web Services -The Web Service Modeling Ontology, Springer Berlin Heidelberg, Germany, 2007. [11] A. Osterwalder, S.B. Lagha, and Y. Pigneur, “An Ontology for Developing e-Business Models”, Proceedings of IFIP DSIAge 2002, Cork, July, 3-7. [12] L. Krstov,and U. Šinkovec, “Relations Between Business Strategy, Business Models and E-Business Applications”, IIS2007,Croatia, September 12-14, 2007. [13] C. J. Anumba, O. O. Ugwu, L. Newnham, and A. Thorpe, “A multi-agent system for distributed collaborative design”, Logistics Information Management, MCB University Press, Vol. 14, No. 5/6, 2001, pp. 355-366. [14] F. Nilsson and V. Darley, “On complex adaptive systems and agent-based modeling for improving decisionmaking in manufacturing and logistics settings”, International Journal of Operations & Production Management, Emerald Group, Vol. 26 No. 12, 2006, pp. 1351-1373. [15] M. Wooldridge and N.R. Jennings, “Intelligent Agents: Theory and Practice”, Submitted to Knowledge Engineering Review, October 1994, Revised January 1995. [16] R. C Nienaber and A. Barnard, “A Generic Agent Framework to Support the Various Software Project Management Processes”, Interdisciplinary Journal of Information, Knowledge, and Management, Vol. 2, 2007. [17] R. O’Connor and J.O.Jenkins, “Using Agents for Distributed Software Project Management”, In Proceedings of the 8th Workshop on Enabling Technologies on Infrastructure for Collaborative Enterprises, IEEE Computer Society, Washington, DC, USA, 1999, pp. 54-60. [18] C. Badica, M. Ganzha, M. Paprzycki, and A. Pirvanescu, “Experimenting with a Multi-agent E-Commerce Environment”, V. Malyshkin (Ed.): PaCT 2005, LNCS 3606, Springer-Verlag Berlin Heidelberg, 2005, pp. 393-402.
[19] Wei-Po Lee, “Towards agent-based decision making in the electronic marketplace: interactive recommendation and automated negotiation”, Expert Systems with Applications, Elsevier Ltd, Vol. 27, 2004, pp. 665-679. [20] S. Razali and M. Osman, “Framework for Agent-Based Buying Decision Process”, Z. Shi and R. Sadananda (Eds.): PRIMA 2006, LNAI 4088, Springer-Verlag Berlin Heidelberg, 2006, pp. 748-753. [21] M. J. Viamonte, I. Praça, C. Ramos, Z. Vale, “Decision Making in an Agent-based Marketplace”, International Conference on Computational Intelligence for Modeling Control and Automation, and International Conference on Intelligent Agents,Web Technologies and Internet Commerce (CIMCA-IAWTIC'06), IEEE Computer Society, 2006.
148
