A Multi-Agent System for Integrating Information from Heterogeneous Data Sources

Abstract

The Internet has changed the way in which information is stored, visualized and shared. In many cases the information we need is on the Internet in different formats and comes from heterogeneous data sources. The tasks of retrieving and integrating information with those features have to be done either manually or by developing a system to do it automatically according to certain rules and criteria. This paper presents a multi-agent system modeled under the BDI approach that integrates information from heterogeneous data sources. Every agent in the group has different abilities and activities, such as retrieving automatically information about call for papers for conferences from different Web sites, which have different structures; extracting information; concentrating it in their own storage system; and visualizing it through a flexible Web site with different search options. Additionally, the multi-agent system provides a Web service interface which allows other systems to retrieve call for papers from it, establishing a certain degree of interoperability to the full system by providing its information in a standard XML format.