A multi-agent system architecture for information gathering

The World Wide Web (WWW) is a vast repository of information. Although there is a great value for this information, its size, dynamic and distribution nature put a significant amount of time and effort to locate, retrieve and integrate the desired information. In addition, different users usually have different views and interest for the same information. The paper proposes a multi-agent architecture that enables cooperation to assist different users to locate, retrieve and integrate information from the WWW. The system architecture comprises three tiers. At the front end, the user agents interact with the users to fulfil their interests and preferences. At the back end, the resource agents access and capture the content and changes of the information resources. At the middle tier, the broker agents facilitate cooperation among the agents to fulfil the user´s desires. A prototype of the proposed system was implemented using the IBM Agent Building Environment Toolkit. The objective is to demonstrate how the agents can cooperate to transparently locate, retrieve and integrate information from dynamic and distributed information resources, utilizing the Internet search engines