SEARCH ENGINE DEVELOPMENT USING EVOLUTIONARY COMPUTATION METHODOLOGIES
Early search engines had their origin in information retrieval systems. These systems were typically developed by human editors to index a document set that was static over long periods of time. The information retrieval systems provided a stable user environment that was eventually optimized over time and facilitative of incremental growth in the document collection. Early search engines used this tried and tested information retrieval model, but encountered usability limitations when the document growth rate accelerated. The limitations of the model became magnified as the need for automated indexing mechanisms grew, and information retrieval systems began to be used with dynamic document datasets. These limitations are still apparent in current search engines which incorporate aspects of these early information retrieval systems. This chapter presents the Tocorime Apicu approach for replacing the information retrieval model with an information sharing model that adapts to changing conditions within the Internet using the stochastic optimization methodologies of evolutionary computation. Experimental results are presented.