A Modified Niche Genetic Algorithm Based on Evolution Gradient and Its Simulation Analysis

To solve the problems of premature convergence and local minima in standard genetic algorithm (SGA), a modified evolutionary gradient-based niche genetic algorithm (GNGA) was proposed. In the GNGA, evolutionary gradient was used to improve the ability of finding the local best; the crossover value and mutation value were adapted dynamically with the generation so that the precision was improved; the population diversity was guaranteed by the use of the niche algorithm based on crowding mechanism. Simulation results show that the proposed algorithm has its superiority in precision and convergence rate compared with SGA.