Automated negotiation in multiple e-Marketplaces by using learning personalized mobile shopping agents

In the B2C- or C2C- based e-Marketplaces, the same merchandise is sold by different sellers in different e-Marketplaces. The trading platforms nowadays has resulted in information explosion that makes buyers unable to retrieve and analyze entire merchandise information easily and, therefore, decreases their negotiation power. Moreover, without the records of buyers' transaction preference, it's not easy for most agent systems to help common buyers to increase their negotiation power. In this paper, we propose the Learning Personalized Mobile Shopping Agent (LPMSA) and apply it to three e-Marketplace architectures: alliance, broker, noncooperation. The buyer can dispatch mobile agents to multiple e-Marketplaces for collecting merchandise information, negotiating with sellers, and buying merchandise from the above architectures. Furthermore, the agent can learn more about buyer's preference. As a result, the proposed architecture can not only find suitable trading partners for buyers but also help them to get better deals.