Towards SOPC architectures for a Complete hardware evolution based Genetic algorithm

Abstract

A Genetic Algorithm GA is a computer based search optimization newlinetechnique that uses the Darwinian Theory of Evolution as a model for newlinefinding exact and approximate solutions GAs belong to a large family of newlineheuristic algorithms called Evolutionary algorithms EA which are being newlineincreasingly utilized for solving complex optimization and search problems newlineThey are basically implemented in either software or in hardware newlineTraditionally GAs are implemented using only software Change being an newlineimportant attribute of GA is handled very easily in software and hardware newlineimplementation was impossible until the advent of reconfigurable hardware newlinetechnology However the large computational time consumed by a GA newlineimplemented in software makes it unsuitable for real time applications newlineToday with the advancements happening in the reconfigurable hardware newlineTechnology this hurdle is overcome thereby shifting the implementation to newlineHardware which drastically speeds up the time factor thus presenting a scope newlinefor real time applications newlineGAs are used for many applications like designing optimization newlinesearch and organization classification and many more Hardware newlinearchitectures in the form of System on Programmable Chip SoPC which are newlinecustomizable for specific GA applications are proposed designed and newlinepresented in this thesis These architectures are based on Complete Hardware newlineEvolution CHE and so the different operations of the GA newline newline