The niching method for obtaining global optima and local optima in multimodal functions
The niching method for obtaining global optima and local optima in multimodal functions
复制标题
多峰函数中求全局最优和局部最优的小生境方法
DOI：
10.1002/scj.10480
复制
发表时间：
2002
期刊：
影响因子：
--
通讯作者：
R. Himeno
中科院分区：
文献类型：
--
作者：
M. Himeno;R. Himeno
有时需要知道多模态函数的二次候选解和全局最优解。确定性拥挤（DC）方法是一种发现多个全局最优解的有效遗传算法，但在发现局部最优解方面存在困难。因此，提出了一种改进的方法，称为“分散确定性拥挤”（DDC），它鼓励个体的分散和种群内物种的创造，以增加对局部最优和全局最优的发现。提出了一种适合度超过某一要求水平的解的优先辨识方法。将DDC算法的性能与其他小生境算法、DC算法、共享算法、RTS算法、带禁忌搜索的遗传算法和免疫算法进行了比较，证明了其有效性。©2003 Wiley期刊公司系统比较，34(11):30 - 42,2003；在线发表于Wiley InterScience （www.interscience.wiley.com）。DOI 10.1002 / scj.10480
Sometimes it is desirable to know the secondary candidates as well as the global optima of multimodal functions. The Deterministic Crowding (DC) method is an effective type of genetic algorithm for discovering multiple global optima, but it has difficulties discovering local optima. An improvement on this method called “Dispersing Deterministic Crowding” (DDC) is therefore proposed which encourages dispersion of individuals and creation of species within the population in order to increase the discovery of local optima, as well as of global optima. A method for preferential identification of solutions with a fitness exceeding some demanded level is also developed. The performance of DDC is compared with those of other niching methods, DC, sharing, RTS, GA with tabu search, and the immune algorithm, to show its effectiveness. © 2003 Wiley Periodicals, Inc. Syst Comp Jpn, 34(11): 30–42, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10480