E-Commerce

 

 

  Home
Applications
Publications

 


    
Semantic Web                Technologies

    MultiAgent Systems
     Electronic Commerce
    Security

     MAS Management

     Information Gathering
    
MultiAgent Learning
    
MAS Interoperability

Software

News
Lectures
Project Grants
Awards
Site Map
People

 

Electronic Commerce (EC) is changing the way business is getting done in the Information Age. To gain a competitive edge, businesses are in need of new computational models and infrastructure. To address this need, we are developing a model of inter-organizational electronic commerce. According to this model, different users are represented by autonomous software agents interconnected via the Internet. The agents act on behalf of their human users/organizations to perform information gathering tasks, such as locating and accessing information from various sources, filtering unwanted information, and providing decision support. Our goal is to coordinate software agents in supply chains dynamically and flexibly so that goods and services can be delivered cost-effectively and when needed.

System Demonstration

As an initial problem domain, we chose collective book purchasing. In the university setting, relatively large numbers of students enrolled in the same classes purchase the same required books. Such groups can be bundled as coalitions, given the ease of payment collection, distribution of goods, and the potentially large number of users involved in the system.

Figure 2:System Architecture
\begin{figure}
\begin{center}
\psfig {file=system2.ps,width=3in}\end{center}\end{figure}

The testbed system (see figure 2) consists of a coalition server, an auctioneer agent, set of supplier agents, and a web-based interface for end users. The system is based on a simple pre-negotiation protocol and a variation of sealed-bid reverse auction that allows suppliers to disclose their discount policies.

Agents use the WWW interface to conduct reverse auctions with supplier agents. The supplier agents, in turn, are given a step function volume discount schedule and make their bids accordingly to projected sizes of coalitions. After the reverse auction is complete, the coalition server opens the coalition to new members, which can join the group if they meet the entrance requirements. After the group is formed, the coalition server proceeds to execute the transaction.

Robotics Institute Project Page

Ecommerce Applications

For more information on our Ecommerce projects, see the following web pages and/or publications:

 

 


Copyright 2006 - 2012 © Advanced Agent-Robotics Technology Lab - The Robotics Institute - Carnegie Mellon University


Internal Site (Restricted Access).