---

Revision with unchanged content. In order to obtain an acceptable quality of filtering services in real time con di tions a trade-off between relevance of results and response time has to be addressed while searching for needed information in numerous distributed re positories. Unfortunately, ignoring a necessity to address information re trie val in distributed and resource-aware manner is usually a serious drawback for many existing search engines that try to survive an ongoing information ex plosion. The essence of a proposed distributed approach for information re trieval is in comprehensive coordination and cooperation mechanisms which both take care of current availability of resources and try to improve them selves during runtime. Additionally, they are able to guarantee delivery of results by combining multiple filtering strategies with very different ca pa bi lities and by using inbuilt self-healing for recovery after failures. As the re a li za tion of the proposed mechanisms is described in details and experimentally pro ven, this book is mainly addressed to professionals in distributed in for ma tion retrieval and advanced search engine technologies. It is also directed to wards researches in filtering strategies and agent technologies.

---

Dr. Dragan Milosevic, information retrieval and machine learning: graduate and master studies of computer science and electrical engineering at University Nis (Serbia) and PhD study of computer science at Technical University Berlin. Vice manager of information retrieval and machine learning competence centre, DAI-Labor, Berlin.

---