EFITA/WCCA 2005

25-28 July 2005, Vila Real, Portugal

e-Negotiations in Agricultural B2B e-Commerce Sotiris T. Karetsos and Constantina I. Costopoulou Informatics Laboratory, Agricultural University of Athens, Iera Odos 75, 11855, Greece, {karetsos, tina}@aua.gr

Abstract Today electronic commerce (e-commerce) has changed the way of doing business, and contributes significantly to economic activity. In any case, e-commerce is not a static field but is always evolving in order to support new and more complex real world processes. Electronic negotiations (e-negotiations) are an important research topic in the area of e-commerce. It ranges from simple offer exchanges to complex communicative acts concerning packages of products and services. In the agricultural sector, enegotiations play an important role in business-to-business (B2B) e-commerce transactions and usually are complicated tasks. B2B partners usually have to negotiate for product quantity, product quality, delivery times and methods, payment methods and price, taking also into account that agricultural products are usually perishable, thus the negotiation period is shortened. This paper deals with the negotiation process in a Multi-Agent Virtual Agricultural Market (MAVAM) which has been build for B2B transactions, using the negotiation agents’ model. An overview of e-negotiations and intelligent agent technology is given and a description of MAVAM system, as well as a specification of MAVAM agents for conducting e-negotiations are presented. Next, Unified Modeling Language (UML) diagrams are produced for the graphical representation of the negotiating agents’ communications, and then the eXtensible Markup Language (XML) is adopted for the standardization of agents’ exchanging messages. Key words: Negotiations, e-commerce, intelligent agents.

1 Introduction Information and communication technologies (ICTs) are rapidly transforming the face of agriculture. Many activities in the agricultural industry are shifting to online environment where interested users (e.g. producers, processors, distributors, wholesalers and customers) are enabling to access a variety of information and services which can be considered as a key element of their agricultural competitiveness. Electronic commerce (e-commerce) can help in this direction. It can be defined as the process of sharing business information maintaining business relationships and conducting business transactions by means of ICTs. Broadly defined, e-commerce is a modern business methodology that addresses the needs of organizations, merchants, and consumers to cut costs, while improving the quality of goods and services and increasing the speed of service delivery (Kalakota, 2000). In the agricultural industry, an increasing number of firms are using various types of e-commerce applications, ranging from simple Web sites to electronic markets. However, the most promising e-commerce applications are found in B2B environment (Costopoulou et al., 2005), where negotiations are an important part of the transaction processes. Negotiation is usually a complicated task in the agricultural sector. This complication arises from the facts that: (i) business partners in the agricultural value chain have to negotiate not just for one but a set of issues such as product quantity, product quality, delivery times and methods, payment methods and price, and (ii) the agricultural products are usually perishable, and thus a short negotiation period is needed. For supporting negotiation processes in different e-commerce applications, several models have been proposed, such as electronic auctions, negotiation agents and negotiation support (Schoop et al., 2003). This paper deals with the negotiation process in a Multi-Agent Virtual Agricultural Market (MAVAM)

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

409

EFITA/WCCA 2005

25-28 July 2005, Vila Real, Portugal

for B2B transactions using a negotiation agents’ model. The rest of the paper is structured as follows: First, an overview of electronic negotiations (e-negotiations) and intelligent agent technology is given. Second, a description of MAVAM system, as well as a specification of MAVAM agents for conducting e-negotiations is presented. Next, a negotiation agents’ model is proposed. Unified Modeling Language (UML) diagrams are produced for the graphical representation of the negotiating agents’ communications, and then the eXtensible Markup Language (XML) is adopted for the standardization of agents’ exchanging messages. Finally, some conclusions are presented.

2 e-Negotiations using Intelligent Agents 2.1 e-Negotiations Negotiation is a decision process in which two or more parties make individual decisions and interact with each other for mutual gain. In tradition, negotiations usually lead to contracts as an outcome which is an agreement between two or more business parties, defining the set of obligations in a business process which reduces uncertainty between the business parties (Chiu, 2003). Negotiation can range over a number of issues such as price, logistics, payment methods etc. as well as from simple to complex, where each participant is an independent decision maker but they all are interdependent because none can achieve goals unilaterally (Kersten, 2003). Thus, with the rapid growth of ICTs, automated negotiation systems became increasingly important in order to support online business processes. The e-negotiation term is considered as the negotiation processes that are fully or partially conducted with the use of ICTs. In other words, negotiators can use electronic media to communicate with each other, determining their contract details. It is estimated that the support of e-negotiations will enhance the ability of a business to maintain large number of partner relationships and also will reduce the need for manual intervention in maintaining those relationships, thereby simplifying life-cycle management of the relationships (OASIS, 2003). According to Scoop et al. (2003), the three main models of e-negotiations are electronic auctions, negotiation agents, and negotiation support. Electronic auctions work according to the general auction principles, i.e. bids concerning defined criteria such as price are placed on a good to be purchased or sold before a predefined end. Negotiation agents take over parts or the whole of the negotiation process for the human principal. Negotiation support approaches do not automate the negotiation process but provide IT support for complex negotiations, leaving the control over the negotiation process to the human negotiators. The agribusiness sector is among those that negotiation processes are a vital component. E-negotiation is considered as a critical element in agricultural e-commerce. While the agricultural products are often perishable, negotiations usually fall into complicated processes and the negotiators have a short period to negotiate. Also agricultural products do not have a standard format and vary in quality and origin thus business partners have to negotiate not just for one but for a set of issues. Shifting to the online environment in order to support agribusiness negotiation process, detailed description of every single step is needed. In this work, we describe the negotiation process for the delivery terms concerning one product. As the number and the allocation of products is increasing the complexity of the negotiating process also rises. 2.2 Intelligent Agents Intelligent agents are a promising technology that defined as software entities that execute functionalities in an autonomous, proactive, social and adaptive fashion. These functionalities include searching, comparing, learning, negotiating and collaborating. Agents have demonstrated their tremendous potential in conducting various tasks in e-commerce such as comparison-shopping, payment, mediation, distribution, sales promotion etc. (Qureshi, 2002). With the increasing importance of e-commerce across the Internet, the need for agents to support business partners in buying and selling goods or services is growing rapidly. In multi agent systems, the participating agents can cooperate to complete a transaction to aid both buyers and sellers. When making a transaction, several agents have to negotiate to achieve a final set of values that satisfies their goal. Thus, a designer of an agent system has the responsibility to determine in detail the agents’ interactions, and also to provide a standard format for encoding messages

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

410

EFITA/WCCA 2005

25-28 July 2005, Vila Real, Portugal

that agents can readily exchange and interpret. In other words, at any point, the agent should know the next step according to the received message from other agents. Therefore, in agent-based systems, enegotiation processes work via the interactive exchange of messages that are proposals and counter proposals. Thus, e-negotiations among agents will consist of numerous messages including terms, explanations, threats, discussions etc. that leads to a total agreement, consensus or a disagreement. Accordingly, a negotiation process among agents consists of the following conversational states (Murthy, 2003): (a) propose: a proposal is being formed and sending to the interested business partner for evaluation, (b) accept: the proposal is accepted, (c) refuse: the proposal is rejected, (d) modify: the business partner based on the initial proposal makes changes and suggests a modified proposal, (e) no proposal: no negotiation, (f) abort: quit negotiation, (g) report agreement: this is the termination point for a successful negotiation, and (h) report failure: this is the termination point for an unsuccessful negotiation. 2.3 eXtensible Markup Language Extensible Markup Language (XML) is a very flexible text format, originally designed to meet the challenges of large-scale electronic publishing. It is also playing an increasingly important role in the exchange of a wide variety of data on the Web and elsewhere. XML has a number of things going for it as a common format for data representation in Internet-enabled B2B e-commerce. It is entirely represented in text which is the primary format used with protocols such as HTTP and SMTP. Therefore XML needs no special handling when used with the most widely employed protocols of the Internet. Because of the technical advantages, XML tools are readily available on most platforms and for most programming languages. The result is that a critical mass of trading partners and suppliers is forming. XML is rapidly transitioning from a technical frontier to the safe choice for data exchange in e-commerce applications. In this context, in order to enable expressive communication in a multi agent system and to organize the communication between agents we use the XML.

3 Negotiation Agents in MAVAM MAVAM is a multi-agent system of a Virtual Agricultural Market (VAM), which has been build for B2B transactions in agricultural markets, providing mechanisms for Internet-based trading and distribution of perishable agricultural products (Costopoulou et al., 2005). The MAVAM actors are: (a) the provider actor who is interested in selling agricultural products using the system, (b) the customer actor who is interested in buying agricultural products using the system, and (c) the transport firm actor who is responsible for delivering the goods after successful matching and negotiation process. In correspondence with MAVAM system’s involved actors, six distinct types of agents have been identified namely, the customer agent, the provider agent, the information brokering and matching agent, the marketing agent, the negotiation agent, and the transport agent. For instance, each customer actor (i.e. buyer or consumer) has his own representative agent called customer agent. This agent is responsible for providing customer information, executing his queries and monitoring the search results. It suggests better search patterns, and tracks the customer’s preferences in the MAVAM system. It is acting autonomously during the registration phase, but when the customer should make a decision for the acceptance or not of an order or negotiate terms, an association between physical customer and his/her software agent is desired. After a successful product matching between the provider agent and the customer agent, the negotiation agent is authorized to negotiate the terms of a business transaction regarding to distribution and transportation terms such as distribution procedures, time, place of delivery, and payment methods. Fig. 1 illustrates the MAVAM system using a use case UML diagram. Based on the MAVAM system, we describe a model for e-negotiation process of B2B transactions between the customer agent and the negotiation agent. As shown in Fig. 1, when the negotiation use case is executed, a negotiation concerning the terms of the physical delivery of the products and the payment method takes place between the customer agent and the transport firm agent. Consequently, during the negotiation between the customer agent and the negotiation agent a set of messages is produced according to the aforementioned conversational states (see sub-section 2.2).

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

411

EFITA/WCCA 2005

25-28 July 2005, Vila Real, Portugal

VAM System

Customer Info

Provider Info

Provider Marketing Info Info Storage (VAM DB)

Customer

Product Order

Information Brokering and Matching

Negotiation

Transport Firm

Transport Order

Fig. 1 MAVAM System The negotiation process has the following phases: (I)

Firstly, the customer agent having found the product to be purchased starts to negotiate the time of delivery, the price and the payment method with a selected transport firm. In order to start the negotiation the transport firm has to know the product type, the product quantity, the customer actor location, the provider actor(s) location, and the delivery time.

(II) Secondly, the negotiation agent (according to scheduled tasks of the transport firm) has the following options: (a) accept: the process is moving to the third phase of negotiation (as described below), (b) refuse-abort-no proposal: the negotiation process is leading to an unsuccessful end and a failure report is issued, and (c) modify: the negotiation agent provides alternative proposal based on the initial one. In that case, the customer agent has to decide, having the following options: (c1) accept: the process is moving to the following phase of negotiation, (c2) refuse-abort-no proposal: The negotiation is leading to an unsuccessful end and a failure report is issued, and (c3) modify: in that case the customer agent makes a new proposal, starting the first phase from the beginning in case that he/she wants to change the place of delivery (e.g. a super market that decides to send this product to another store) and then the negotiation agent evaluates the new proposal. (III) Thirdly, after a successful second phase the negotiators are about to start negotiating about the price and the payment method. The negotiation agent proposes the cost estimation and the preferred payment method to the customer agent. The customer agent has to decide having the following options: (a) accept: the negotiation ends successfully and an agreement report is issued, (b) refuseabort-no proposal: the negotiation is leading to an unsuccessful end and a failure report is issued, and (c) modify: the customer agent provides an alternative proposal, based on the initial. In this case, the negotiation agent is about to decide, and also has the same options: (c1) accept: the negotiation ends successfully and an agreement report is issued, (c2) refuse-abort-no proposal: The negotiation is leading to an unsuccessful end and a failure report is issued, and (c3) modify: in that case the

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

412

EFITA/WCCA 2005

25-28 July 2005, Vila Real, Portugal

customer agent evaluates the new proposal (starting again the third phase). A more detailed description of the aforementioned procedure, involving the customer agent and the negotiation agent, is given using a UML activity diagram (Fig. 2). This diagram is analyzing the customer and the negotiation agents’ decision process. Also, the XML formation (Table 1) of the exchanging messages is given which describes the first phase of negotiation. More specifically, the XML formation describes a real-life scenario about a super market which performs tasks via the customer agent that initially wants to purchase oranges from Portugal and to deliver to Greece in 25th of July 2005, and the negotiation agent which acts on behalf of the transport firm is modifying this proposal for 29th of July of 2005. Next, the customer agent modifies the initial proposal by changing the place of delivery for a store in Cyprus for the same date, which is accepted from the negotiation agent.

Fig.2 Negotiation process expressed as an activity diagram

Table 1: Agent communication XML messages 1. MESSAGE FROM Customer Agent Super Market A 75 Iera Odos Athens Greece 11855 EFGH 120 Real Street Vila Real

Portugal 12345 Oranges 2000 kgr 10 kgr 2005-07-25

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

413

EFITA/WCCA 2005

2. MESSAGE FROM Negotiation Agent 2005-07-29 3. MESSAGE FROM Customer Agent Super Market A 10 Nowhere Street Nicosia Cyprus 56789 EFGH 120 Real Street

25-28 July 2005, Vila Real, Portugal

Vila Real Portugal 12345 Oranges 2000 kgr 10 kgr 2005-07-25 4. MESSAGE FROM Negotiation Agent Accepted

4 Conclusion In this work, we discussed an e-negotiation process model in a multi-agent system for B2B transactions in agricultural markets. The agent technology can assist to the automation of such business processes, acting on behalf of their users. Agents are able to discuss and cooperate in such a way that gives meaning to their messages, so as to make comprehensive dialogues. For this reason, we take advantage of the capabilities of XML for the formation of the interchanging messages. In a subsequent phase, we intend to proceed to the description of more complicated processes concerning a number of products and multiple places of delivery.

5 References Chiu, D.K.W., Cheung, S.C., Hung, P.C.K., 2003. Developing e-negotiation process support by Web services. Proceedings of the International Conference on Web Services. Costopoulou, C.ǿ., Lambrou, M.ǹ., Karetsos, S.T., 2005. A multi-agent system for Internet middlemen in B2B environment: A case study in agribusiness. Journal of Applied Systems Studies. Cambridge International Science Publishing. Cambridge, England (in press). Kalakota, R., Whinston, A. B., (Eds.), 2000. E-Business 2.0 Roadmap for success. Addison-Wesley Publishing Company. Kersten, G.E., 2003. The Science and Engineering of E-Negotiation: An Introduction. Proceedings of the 36th Hawaii Conference on System Sciences (HICSS’03). Murthy, V.K., 2003. Designing Agent-Based Negotiation for E-Marketing. In: Nansi, S., Murthy, V.K., (Eds.), Architectural Issues of Web Enabled Electronic Business. United Kingdom, pp. 292-304 OASIS ebXML Collaboration Protocol Profile and Agreement Technical Committee, 2003. Automated Negotiation of Collaboration-Protocol Agreements Specification. Version 0.01. http://lists.oasis-open.org/archives/ebxml-cppa-negot/200210/pdf00000.pdf Qureshi, L., 2002. E-Negotiations Using Intelligent Agents. Department of Information Systems and Computing, Brunel University (Dissertation). Schoop, M., Jertila, A., List, T., 2003. Negoisst: a negotiation support system for electronic business-tobusiness negotiations in e-commerce. Data & Knowledge Engineering. Vol. 47, pp. 371-401.

2005 EFITA/WCCA JOINT CONGRESS ON IT IN AGRICULTURE

414