New gender genetic algorithm for solving graph partitioning problems

An efficient genetic algorithm based on gender selection is proposed. In this algorithm, each individual (chromosome) in the population has an additional feature, its gender. Individuals are arranged in descending order according to their fitness values, then a gender is assigned to each chromosome by simply alternating female with male. Crossover, or mating, is only permitted between individuals of opposite genders. The new gender-based genetic algorithm (GGA) is applied to partitioning problems and its performance is compared to the conventional genetic algorithm (GA). Experimental results indicate that GGA significantly outperformed GA in terms of the quality of the optimum solution and the number of generations (convergence).