A framework for applying intelligent agents to support electronic trading
Introduction
Electronic commerce (EC) is the cutting edge for today's business. The widespread of Internet opens an enormous amount of business opportunities. More and more organizations are facing the challenge of this new technology [8], [10], [16], [31]. There are several motivations for doing business electronically. First, Internet users have become a fast growing group that forms a promising market. A report estimated that there were more than 28 million Internet users up to 1996 in the US alone [11]. America Online (AOL) alone is claiming to have 15 million users in 1998. The market size is estimated to be tens of billions by the year of 2000 [18]. This is attractive to virtually any business. Second, Internet offers a new way of doing business, which can overcome time and geographic barriers. Traditional businesses have specific business hours and can only serve customers within a geographic range. Since Internet links customers all around the world, customers can purchase virtually any products from anywhere in the world as long as they are available on the Internet. This will have a tremendous impact on traditional stores [3], [24]. Third, EC allows service providers to have more information about their customers and customers are also likely to receive a better service. For instance, sellers can collect customer information through the network, whereas customers can easily obtain better products or services by accessing more stores or choosing from more available alternatives. New business ideas and channel revolution are also underway [3].
Although EC is attractive; its transaction process is often complicated. Involved parties may need to collect and analyze information, negotiate contracts, execute transactions safely, and provide follow-up services over the Internet. Therefore, it is critical to develop environments that can handle the growth of electronic markets and control the increase in complexity. In fact, controlling information overload in EC may be a key to its future.
One way to reduce information overload is to delegate some activities to software agents. An agent is a person or business authorized to act on another's behalf. In software development, an agent is a computer program that can operate autonomously and accomplish unique tasks without direct human supervision [32]. A software agent possesses the properties of autonomy, social ability, reactivity, and pro-activeness [30]. It is often used to manage information, support decision-making, and automate repetitive office and personal activities on behalf of the user. In some cases, an agent needs internal knowledge to perform the task intelligently. This is called an Intelligent Agent (IA). The agent approach has been applied to e-mail filtering [22], simulation [28], learning [25], programming [27] and many other domains. Therefore, it is reasonable to anticipate that, if properly applied, IAs can effectively reduce the load on the user and hence increase the performance of electronic transactions.
The purpose of this paper is to study the role of IAs in EC and to develop a framework for applying IAs to support business activities in electronic environments. The remainder of the paper is organized as follows. First, literature in EC and IAs are reviewed to give an overview of the roles, transaction mechanisms, and existing applications of IAs in EC. Then, activities involved in electronic transactions are analyzed and classified. An agent-based framework for EC is presented. It groups agents into three levels: market, contract, and activity. Agents at a lower level support those at the higher level to achieve a high integrity. Finally, sample applications of the framework are illustrated.
Section snippets
Literature review
In order to know how IAs can be applied to support EC, we need to know different types of electronic transactions and their respective IA applications.
Analysis of trading activities
In most commercial process, there are three major players: buyer, vendor, and broker. Buyers are customers who purchase certain products or services. Vendors are product or service providers. Brokers are intermediaries who help the buyer and the vendor to complete a transaction. The buyer and vendor must exist in any trading, while the broker exists only in certain conditions. These players form different trade types. Each type has a set of activities to be performed. A good environment for EC
Framework for EC agents
Given different trade types, applying IAs can play a significant role in EC. For instance, a bidding agent can help call for bidding, solicit bids, notify winners, and perform transactions over the Web. In fact, applying IAs can also change traditional trade types. For example, agents serving as brokers on the Web may add a third party to bilateral transactions.
The discussion in Section 3 indicates that IAs can be used to support EC at three different levels: market, contract, and activity, as
Illustrative examples
To demonstrate the above agent-based EC environment in detail, an illustrative example is presented in this section. For agent communication, we adopt the Knowledge Query and Manipulation Language (KQML) [9] as the outer language, and an inner language similar to UNIK-OBJECT [12] to describe contracts. Three most commonly used commands (called performative in KQML) are ask-if, evaluate, and reply.
ask-if
:reply-with <expression>
:sender <word>
:receiver <word>
:content <expression>
Ask-if is a
Concluding remarks
EC is becoming an important channel for future business. In this paper, we have presented a three-level framework for using IAs to support EC. It groups agents into three layers: market, transaction, and activity. Agents at the market level accept requests from the user and choose a proper trade type. Agents at the transaction level ensure that the selected type is executed properly. Agents at the activity level perform a specific task in the user's decision making process. Because the
Acknowledgements
The authors thank the reviewers for their comments on earlier versions of the manuscript. The research was partially funded by a grant from National Science Council to the first author.
Ting-Peng Liang is Professor of Information Systems and Director of the Software Incubator of the National Sun Yat-sen University in Kaohsiung, Taiwan. Prior to the current position, he was the Director of the Graduate Institute of Information Management and Dean of the College of Management. His research interests include electronic commerce, decision support systems, model management, and intelligent systems. His papers have appeared in journals such as Management Science, Operations
References (32)
- et al.
Using client-broker-server architecture for intranet decision support
Decision Support Systems
(1997) - et al.
Key issues in information systems management: a Delphi study in Slovenia
Information and Management
(1996) - et al.
Requirements for the design of a protocol suite to automate DSS deployment on the World Wide Web: a client/server approach
Decision Support Systems
(1997) - et al.
An empirical study on consumer acceptance of products on electronic market: a transaction cost model
Decision Support Systems
(1998) Key information management issues in Taiwan and the US
Information and Management
(1996)- et al.
Electronic markets and virtual value chains on the information superhighway
Sloan Management Review
(1995) - et al.
The World Wide Web as an advertising medium: toward an understanding of conversion efficiency
Journal of Advertising Research
(1996) - et al.
Decision support on demand: emerging electronic markets for decision technologies
Decision Support Systems
(1997) - et al.
Scalable internet resource discovery: research problems and approaches
Communication of the ACM
(1994) - et al.
Intelligent spider for Internet searching
Information retrieval in the World Wide Web: making client-based searching feasible
Internet and Web use in the US
Communication of the ACM
Marketing in hyper- media computer-mediated environments: conceptual foundations
Journal of Marketing
Intelligent agents for matching information providers and consumers on the World Wide Web
Intelligent agent based contract process in electronic commerce: UNIK-AGENT approach
Cited by (71)
VCAs as partners or servants? The effects of information sensitivity and anthropomorphism roles on privacy concerns
2023, Technological Forecasting and Social ChangeLocal matching of flexible load in smart grids
2016, European Journal of Operational ResearchCitation Excerpt :Moreover, the generality of the model allows diverse demand such as energy storage, data centers, HVAC, agricultural pumping or others to be described within this model. Naturally, individual agents may require specific decision support to map their flexibility endowments into the bidding space (Liang & Huang, 2000; Reeves, Wellman, MacKie-Mason, & Osepayshvili, 2005). Again we want to reiterate the idea of commoditization which does not allow for highly differentiated market products accounting for heterogeneous ramping rates or availability agreements.
Multi-agent technology and ontologies to support personalization in B2C E-Commerce
2014, Electronic Commerce Research and ApplicationsCitation Excerpt :Following a process flow formed out of six stages (quite like the CBB model), a customer is monitored and supported in the B2C activities. Another framework is proposed in (Liang and Huang 2000) to coordinate EC activities, by organizing agents in three layers (market, transaction, activity). Market Maker agents (belonging to the market layer) coordinate tasks.
A Web-based personalized recommendation system for mobile phone selection: Design, implementation, and evaluation
2010, Expert Systems with ApplicationsCitation Excerpt :Overall, we show that the AHP method can be used as a core algorithm for automated recommendations in complex product search spaces. Our study also has implications for designing and using intelligent e-commerce software agents that offer autonomy, reactivity, and proactiveness (Wooldridge & Jennings, 1994) to support consumers’ decision making (Liang & Huang, 2000). For example, an (intelligent) agent could automatically gather essential product information from identified Web sites and, using preferences previously collected from the consumer, identify products appropriate for him or her.
SIMBA: A simulator for business education and research
2010, Decision Support Systems
Ting-Peng Liang is Professor of Information Systems and Director of the Software Incubator of the National Sun Yat-sen University in Kaohsiung, Taiwan. Prior to the current position, he was the Director of the Graduate Institute of Information Management and Dean of the College of Management. His research interests include electronic commerce, decision support systems, model management, and intelligent systems. His papers have appeared in journals such as Management Science, Operations Research, Decision Sciences, Decision Support Systems, MIS Quarterly, Journal of MIS, among others. He is also on the editorial boards of many academic journals.
Jin-Shiang Huang is an Assistant Professor of Information Systems at Ming Chuan University in Taipei, Taiwan. He received his MBA and PhD degrees in Information Management from National Sun Yat-sen University. His research interests include economics of electronic commerce, decision support systems, and negotiation support systems. His papers have appeared in Decision Support Systems and conference proceedings.