Presents a theoretical framework for building intelligent agents to gather information from the Web. The thesis presents both theoretical and technical solutions to the main problems of this area. Such kinds of agents can gather relevant information from highly distributed collections to satisfy user information needs.