Playing the Web Search Game∗

Distributed heterogeneous search environments are an emerging phenomenon in Web search. Together with new promises they bring new challenges. We describe COUGAR, a learning system that competes in such environments. A heterogeneous search environment consists of multiple specialised search engines and metasearchers. All these components can be independently owned and controlled. Specialised search engines provide focused search services in a specific domain (e.g. on a particular topic). Different search engines can specialise in different domains (hence, the heterogeneous environment). Metasearchers help to process user queries effectively and efficiently by distributing them only to the most suitable search engines for each query. A typical scenario works as follows. To find a “suitable” search engine, users submit search requests to the metasearcher. The metasearcher has a database with index content summaries of known search engines. For each search query, the metasearcher selects the search engine(s) that provides the best quality (most relevant) results. Fig. 1 illustrates the interactions in a heterogeneous search environment.