A hybrid strategy for differential variation
Xiaowei Zhang · Caai Transactions on Intelligent Systems · 2008
A new hybrid strategy for differential variation was developed in order to improve the performance of differential evolution.The strategy regards each individual in the population as a charged particle,and uses the charge level of the particle and the attraction-repulsion mechanism among particles to determine the probable direction and magnitude of motion.This strategy can compel individuals to move adaptively in the direction of the forces exerted on them by three other individuals.Numerical experiments showed that differential evolution based on this strategy has high accuracy and requires fewer evaluations.