A Modified Genetic Algorithm with Maintaining Diversity
Zhang Cui-fang · Acta Simulata Systematica Sinica · 2005
This paper presents a modified genetic algorithm with maintaining diversity for solving genetic excursion and premature convergence. In the algorithm, entropy of population and locus diversity take as the measure of population diversity of the evolving state. The relation between genetic operation and population diversity is set by their function, which makes the main parameter of genetic operation adaptively adjust according to the varying fitness function and diversity function. The algorithm greatly improves population diversity and the searching efficiency. The experiments show advantages of the algorithm.