Agent Models in ECommerce

Software Agents have become central to the success of second-generation e-commerce that calls for increased levels of personalization, context sensitivity and dynamism. While agents can play a variety of roles in facilitating online trading, this work chooses to focus on two specific roles viz. the role of agents in Product Brokering / Recommendation and the role of agents in Negotiation. The architecture of the recommendation agent presented in the former part, utilizes a novel algorithm called Entropy based Collaborative Filtering to predict users’ ratings to individual items and then provide them with a personalized set of recommendations. The latter part develops an automated negotiation procedure for bilateral multi-issue negotiations under two-sided information uncertainty by addressing the mechanism design and agent design aspects in an integrated manner. The novel ideas presented for both product recommendation and automated negotiation shall appeal researchers owing to their theoretical backing while they shall appeal e-commerce practitioners owing to their simplicity, and implementation feasibility.