Some new mutation operators for genetic data clustering

Genetic algorithm is one of evolutionary algorithms which have been used widely to solve many problems such as data clustering. There are lots of genetic data clustering algorithms which have worked on fitness function to improve the accuracy of algorithm in evaluation of generated chromosomes and have used simple and all purpose crossover and mutation operators such as one point crossover and random change mutation. Mutation process randomly modifies the gene values at selected locations to increase genetic diversity, by forcing the algorithm to search areas other than the current area. Simple non heuristic mutations such as random change mutation increase genetic diversity but they also increase execution time and decrease fitness of population. In this paper we introduce some new heuristic mutation operators for genetic data clustering. Experimental results show that all of proposed mutation operators creates better offspring than random change mutation and increases the fitness of population.