Personalized multimedia news agents in a broadband environment.

Abstract

The growing demand for efficient, flexible yet cost effective services has produced a new perspective on personalized service as a form of asynchronous interactive computing with seamless support for communications and network. Personalization of services exists at two levels: the user and remote service provider levels. At the user level is a customized interface that provides access to networked services unlimited by distance. At the remote service provider level, the aim is to offer services based on individual needs as expressed in the user customized interface. The gap between the two sides is filled by the activities of agents. Software agents are entities that constantly adapt to the needs of the user in terms of interests, time and space limitations. Agents observe and learn from the user the art of selecting information of value to the user. Autonomous agents have the right to make decisions on behalf of the user when dealing with other agents, resources or service providers. The processing involved can be done either locally (when static) or remotely (when mobile). This thesis presents the architecture and environment of an agent-based system. The design of news agents for personalized news delivery over a broadhand network is proposed. The mobile-agent paradigm has been derived from the demands of easy remote services accessibility. A new model, namely aclient-iserver model, has been created to provide an interactive and asynchronous solution to the existence of mobile agents. Personalization of services is achieved by collecting user interests locally and gathering information remotely, all transparent to the user.