طرح مذاکره مبتنی بر عامل چند مسئله ای دو جانبه برای انجام معاملات الکترونیکی بازار

Software agents are flexible, autonomous, and dynamic computational entities. For electronic market (e-market) applications, the wide variety of choices to the consumers has introduced the problem of information searching. Meanwhile, there are also so many issues for the participants that require negotiations to find the best deal. In this paper, we present a bilateral multi-issue negotiation scheme, which deploys risk seeking/averse strategies and cost-oriented/benefit-oriented tactics to explore the negotiation space of possible trade-off for the one that is most likely to be acceptable. In particular, it focuses on the incorporation of the level of vagueness of preferences for those issues (either crisp or fuzzy) and multi-objective decision-making techniques. The goal of this research is to design an effective model and some constraints-based fuzzy rules and relax of these constraints to their prevailing transaction circumstances.

Recently the rapid proliferation of the Internet is stimulating the transformation of traditional markets into electronic markets (e-markets). Classical theories cannot adequately cover the dynamics of complex computing systems, such as the e-market, creating the need for new methods. Consequently, computing scientists and experts have joined their efforts with economists and mathematicians to create a novel approach for modeling and evaluating market processes. Software agents are encapsulated computer systems situated in some environments such as the Internet and are capable of flexible, autonomous actions in e-market environment to meet their design objectives [1] and [2]. It is argued that software agents can provide a high level of intelligence and autonomy for enhancing the effectiveness of e-market. For example, a person that wants to purchase an item can be represented by a software agent. This agent negotiates for the item on behalf of the person and can eventually purchase the item for him when all conditions are met. Software agents can, meanwhile, incorporate experiential knowledge of past transactions to streamline the effects of volatile demand across multiple e-marketplaces. The negotiating agents are capable of exchanging proposals, evaluating proposal, and also accepting or rejecting proposals to reach mutual deals. Thus, there is a growing importance of understanding how people develop software agents for these fields [3]. Software agents are autonomous, interact with other agents, and enable approaching inherently distributed problems with negotiation and coordination capabilities [4]. The negotiation mechanism is often the most complex, since it requires evaluation and decision-making under uncertainty, based on multiple issues (attribute) of quantitative and qualitative nature, involving temporal and resource constraints, risk and commitment problems, varying tactics and strategies, domain specific knowledge and information asymmetries, etc. [5]. Thus, the objective of this paper is to develop the agent-based negotiation scheme which can be seen as a decision-making process of automatically resolving a conflict involving many parties over mutual goals to enhance negotiation agents’ decision-making processes in e-market. In this research, we propose a series of negotiation processes through employing fuzzy logic to evaluate different scale of each issue, generating similarity matching with bilateral alternatives offered by buyer agent and supplier agents, and then modeling some constraints-based fuzzy rules for sellers when receiving counter-proposal by buyer, consequently proceeding to trade-off mechanism between both sides to gain an agreement. This negotiation scheme is mainly classified into five parts. We first defined negotiation parameters set and iso-curve computation in preliminary setting. Second, negotiation alternative processing service will be proposed to select buyer's alternative based on iso-curve. After selecting negotiation alternative, buyer agent will send its alternative (counter-proposal) to supplier agents to determine if it satisfies suppliers’ constraints or not, then decide iso-curve relaxation which is buyer's subjective behavior. Consequently, we use trade-off to find out buyer's partner and determine which issues need to change which along with the iso-curve. At the last part, post-negotiation analysis, we will compare two results from the preceding trade-off strategies (i.e., risk-seeking and risk-averse) to let decision-making know how to make the most beneficial decision for company. An example application to negotiate a supplier selection among participants demonstrates to the agents how to adjust its negotiation issues parameters and reach its agreement. In this paper, our main contribution illustrates the design and development of negotiation agents to enhance the degree of autonomy and the efficiency of e-market. In particular, it is providing a platform which can negotiate not only quantitative attributes but also qualitative issues realized fuzzy logic. And make this negotiation scheme more practical in a real world and worthy to implement. The rest of the paper is organized as follows. Section 2 briefly reviews some methodologies proposed of negotiation. The details of the scheme we proposed and the reasons that lie behind are given in Section 3. The performances of the one-to-many multi-issue negotiation scheme are evaluated and discussed in Section 4. The application is explained through the computers assembling e-market as our case study in Section 5. Additionally, the novel features of agents in this work are compared with several related systems in Section 6. Finally, the last section contains some conclusions and perspectives.

Negotiation recently was developing dramatically and very important to the success for a e-commerce. This is because the cost and quality of goods and services are usually concerned by buyer and need to negotiate. Therefore, we developed an automated bilateral multi-issue negotiation scheme for facilitating software agents technology approaches in an outsourcing partner selection case. Suppliers and buyer in agent-based e-Market through negotiations have significant impact on supplier selection and partners’ profit. Using the fuzzy value in fuzzy logic transfers linguistic value into the same scale in order to evaluate sequent process. Then through the transaction between buyer's and suppliers’ agents to let both sides know opponents’ behavior and decide which trade-off strategy can be implied. Among the negotiation, similarity matching applied in two parts; one is to find the most similar alternative as the counter-proposal of buyer, the other is for buyer to choose a business partner and know which issues has to change and make a consensus with suppliers. This negotiation contributed more practical approach; because it included fuzzy logic to represent the issues and jointed buyer's behavior within negotiation, therefore, this research makes close to negotiate in a real world. In the future work, we will make more efforts on post-negotiation to let decision-maker have more accurate decision.