A clustering algorithm using genetic algorithm with competitive individuals

We propose a clustering method for profile data based on a genetic algorithm (GA). The GA can be considered as a flexible algorithm by setting up genetic operations and fitness function according to the problem. We propose the competitive GA considering direct competitions between individuals concerning the representative patterns for clusters. Further, we tried some specific genetic operations for clustering problems. The efficiency of this algorithm was examined in the clustering problem of the fluorescence profiles of chromosomes which are measured by digitizing the fluorescence intensities emitted from slit-scanned chromosomes.