Dynamic load balancing model: Preliminary assessment of a biological model for a Pseudo-Search Engine

Abstract

Emulation of the current World Wide Web (WWW) search engines using methodologies derived from Genetic Programming (GP) and Knowledge Discovery in Databases (KDD) were used for the Pseudo- Search Engine's initial parallel implementation of an indexer simulator. The indexer was implemented to follow some of the characteristics currently implemented by AltaVista and Inktomi search engines who index each word in a Web document. This approach has provided very thorough and comprehensive search engine results that have led to the development of a Pseudo-Search Engine Indexer which has in turn provided insight into the computational effort needed to develop and implement an integrated search engine - information crucial to the adaptation of a biological model. The initial implementation of the Pseudo-Search Engine Indexer simulator used the Message Passing Interface (MPI) on a network of SUN workstations and an IBM SP2 computer system. © 2000 Springer-Verlag Berlin Heidelberg.