自适应代表个体选择方法的协作协进化算法

协作协进化算法中,代表个体选择按贪心度不同可分为最优选择和随机选择.最优选择对于大多数子模块之间关联性不是很强的问题都较为有效,但对于子模块间有很强关联性的问题,随机选择比最优选择更为有效.所以,提出一种自适应代表个体选择方法的协作协进化算法,该算法通过进化停滞判断函数将最优选择和随机选择两种代表个体选择方法结合起来.仿真结果表明,该算法对于子模块间强弱不同的问题都能有效的找到解.

Adaptive representation selection of cooperative co-evolution Algorithm

With the different degree of greediness, there are two methods of representation selection in cooperative co-evolution algorithm: best selection and random selection. Best selection can do well in most situation when the inter-activity between sub-components of the problem is not so strong. But to the problem with strong inter-activity between sub-components, random selection is better than best selection. In this paper we proposed an adaptive representation selection cooperative co-evolution algorithm. This approach combines two representation selection ways of best selection and random selection, by a stagnate judge function. The simulation results show that this algorithm is effective to the problems with different inter-activity degree between sub-components.