文献类型：期刊文献

中文题名：Gaussian process assisted coevolutionary estimation of distribution algorithm for computationally expensive problems

英文题名：Gaussian process assisted coevolutionary estimation of distribution algorithm for computationally expensive problems

作者：Luo Na[1];Qian Feng[1];Zhao Liang[1];Zhong Wei-min[1]

机构：[1]E China Univ Sci & Technol, Key Lab Adv Control & Optimizat Chem Proc, Minist Educ, Shanghai 200237, Peoples R China

年份：2012

卷号：19

期号：2

起止页码：443

中文期刊名：Journal of Central South University

外文期刊名：JOURNAL OF CENTRAL SOUTH UNIVERSITY

收录：;EI(收录号：20122615184623);Scopus;WOS:【SCI-EXPANDED(收录号:WOS:000299928600020)】；

基金：Foundation item: Project(2009CB320603) supported by the National Basic Research Program of China; Project(IRT0712) supported by Program for Changjiang Scholars and Innovative Research Team in University; Project(B504) supported by the Shanghai Leading Academic Discipline Program; Project(61174118) supported by the National Natural Science Foundation of China

语种：英文

中文关键词：estimation of distribution algorithm; fitness function modeling; Gaussian process; surrogate approach

外文关键词：estimation of distribution algorithm; fitness function modeling; Gaussian process; surrogate approach

摘要：In order to reduce the computation of complex problems, a new surrogate-assisted estimation of distribution algorithm with Gaussian process was proposed. Coevolution was used in dual populations which evolved in parallel. The search space was projected into multiple subspaces and searched by sub-populations. Also, the whole space was exploited by the other population which exchanges information with the sub-populations. In order to make the evolutionary course efficient, multivariate Gaussian model and Gaussian mixture model were used in both populations separately to estimate the distribution of individuals and reproduce new generations. For the surrogate model, Gaussian process was combined with the algorithm which predicted variance of the predictions. The results on six benchmark functions show that the new algorithm performs better than other surrogate-model based algorithms and the computation complexity is only 10% of the original estimation of distribution algorithm.
In order to reduce the computation of complex problems, a new surrogate-assisted estimation of distribution algorithm with Gaussian process was proposed. Coevolution was used in dual populations which evolved in parallel. The search space was projected into multiple subspaces and searched by sub-populations. Also, the whole space was exploited by the other population which exchanges information with the sub-populations. In order to make the evolutionary course efficient, multivariate Gaussian model and Gaussian mixture model were used in both populations separately to estimate the distribution of individuals and reproduce new generations. For the surrogate model, Gaussian process was combined with the algorithm which predicted variance of the predictions. The results on six benchmark functions show that the new algorithm performs better than other surrogate-model based algorithms and the computation complexity is only 10% of the original estimation of distribution algorithm.