一种改进的多种群遗传算法研究

In older to combat premature convergence inherent in simple genetic algorithm (SGA), an improved multi-population genetic algorithm (IMPGA) is proposed. It evaluates the species using unison fitness, and proposes a local competition method to keep the elitist individuals. The paper also proposes the disturbance strategy to improve premature convergence, so that the algorithm can get away from the local optimal value. The experimental results of this method on the classical complex function show the algorithm has a strong ability to get away from the local optimal value.