Skip to Main Content
As the volume and variety of information sources continues to grow, especially on the World Wide Web (WWW), the requirements imposed on search applications are steadily increasing. The amount of available data is growing and so do user demands. Users do not need more information to deal with. Rather, they require that the information search process provides them with sensible responses to their requests. There are several problems complicating the search process and lowering the search effectiveness: users rarely present search queries in the form that optimally represents their information needs; the measure of a document's relevance is often highly subjective among different users; and information sources contain heterogeneous documents, stored in multiple formats and without a standardized representation. To alleviate these problems, queries can be extended using the concepts of fuzzy sets. The search system described in this paper models users' information needs in a framework of fuzzy sets, with the aid of two metrics that determine the "fitness for use" of the retrieved documents. With the aid of this parameter, an evolutionary computing system performs optimization of search queries based on individual user models. This way, an effective search system is produced, which is able to continuously learn from reinforcements provided by the users.