摘要
协同进化算法中,计算个体适应度时,代表个体的选择以及代表个体与个体的组合评估需要很大的计算量。协同进化遗传算法虽然计算量相对小一点,但是只能获得一个贪婪解。多模式共生进化算法虽能克服协同进化遗传算法的这个缺点,但是计算量太大。本文利用间隔时间学习方法提出间隔时间学习协同进化算法,该算法每隔N代交互一次信息。在此基础上,将抽样法应用到协同进化算法中。实验结果表明,这种方法能有效地减少计算量,且本文从数学方面进行了分析验证。
When evaluating individuals, the selection of representation and the evaluation of the combination of the individuals and representations need lots of computation in co-evolution algorithm. For cooperative co-evolutionary genetic algorithm, calculated amount is small, but it can only obtain one greedy solution. Multi-pattern symbiotic evolutionary algorithm can overcome the shortcoming, but its calculated amount is too big. In this paper we proposed punctuated anytime learning co-evolution algorithm using punctuated anytime learning method, this approach interacts information every N generations. Based on this algorithm, the sampling method was used to co-evolutionary algorithms. The experimental results and mathematical analysis show that this algorithm is effective to reduce the calculated amount.
出处
《电子测试》
2012年第3期16-19,共4页
Electronic Test
关键词
协同进化算法
协同进化遗传算法
多模式共生进化算法
间隔时间学习
抽样法
co-evolution algorithm
cooperative co-evolutionary genetic algorithm
multi-pattern symbioticevolutionary algorithm
punctuated anytime learning
sampling method
