Supporting Electronic Negotiation for Intelligent Trading

Abstract

Intelligent negotiation agents are software agents, which can negotiate the terms of transactions on behalf of purchasers and vendors on the Internet. Current solutions are mostly limited to single attribute negotiations, and are typically used to determine price. Moreover they typically assume information to be precisely defined and shared between the parties. Bargaining situations are, in most cases, characterized by conflicting interests among the agents that don’t cater for common interests and possibility for collaboration to improve the outcomes of the parties. Another limitation of existing on-line negotiation agents is that their negotiation is usually taking place in a centralized marketplace where the agents meet and negotiate following a set of protocols that don’t cater for more open and direct party-to-party negotiations. This chapter reports on solutions for addressing the issues of negotiations with incomplete and imprecise information, dynamic coalition formation and negotiation ontologies. The negotiation with incomplete and imprecise information uses fuzzy constraint-based reasoning and the principle of utility theory. The formation of coalition is based on negotiation over the distribution of the coalition value and the agent level of resources. The negotiation ontologies make use of shared ontologies as well as individual ontologies to avoid misunderstanding and make data exchange meaningful.