A Study of Selection Operator Using Correlation Between Individuals in Genetic Algorithm

It is one of the important problems to maintain the diversity in genetic algorithm. Generally, when the diversity is not fully maintained, it is difficult to find good solution. In this paper, we propose an elite correlation selection operator as a new selection operator to improve the genotype diversity in genetic search process. The proposed method selects individuals with a fitness and a genotype correlation between individuals. As a measure of the correlation, we use the hamming distance of the genotype. The performance of the proposed method is examined with variable knapsack problem. We compare the performance of the proposed method with conventional methods. The experimental results show good performance.