Pipelined processing of genetic clustering

In this article, an attempt is made to solve the clustering problem using a pipelined genetic algorithm, which is a faster version of the conventional genetic algorithm. In the pipelined genetic algorithm, the genetic operators are pipelined in such a way that their execution overlaps in time; thereby providing an overall speedup of the process. Since the conventional selection requires the computation of the fitness values of all the chromosomes before the selection process begins, it has been modified in pipelined GA to a version that uses a statistical distribution function. Both crisp and fuzzy clustering problems can be solved using pipelined GA. An analysis of the speedup obtained in case of software pipeline is provided.