VEMA: Multi-agent system for Electronic Commerce on Internet

With the increasing importance of EC(Electronic Commerce) on the Internet, the need for intelligent software agents to support both customers and suppliers is growing rapidly. The lack of structure of information and negotiation makes full automation of EC infeasible. We focus on an adaptive personalized agent and open infrastructure for agent-based EC. For an adaptive personalized agent, we employ hybrid learning mechanism to learn user model in the EC domain. As an open infrastructure, we propose a complete architecture, and message and protocol for negotiation between agents.