Unified eigen analysis on multivariate Gaussian based estimation of distribution algorithms

被引:54
作者
Dong, Weishan [1 ]
Yao, Xin [2 ]
机构
[1] Chinese Acad Sci, Inst Automat, Key Lab Complex Syst & Intelligence Sci, Beijing 100190, Peoples R China
[2] Univ Birmingham, Sch Comp Sci, CERCIA, Birmingham B15 2TT, W Midlands, England
基金
英国工程与自然科学研究理事会;
关键词
estimation of distribution algorithm; eigen analysis; multivariate Gaussian distribution; covariance matrix scaling; eigenvalue tuning;
D O I
10.1016/j.ins.2008.01.021
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Multivariate Gaussian models are widely adopted in continuous estimation of distribution algorithms (EDAs), and covariance matrix plays the essential role in guiding the evolution. In this paper, we propose a new framework for multivariate Gaussian based EDAs (MGEDAs), named eigen decomposition EDA (ED-EDA). Unlike classical EDAs, ED-EDA focuses on eigen analysis of the covariance matrix, and it explicitly tunes the eigenvalues. All existing MGEDAs can be unified within our ED-EDA framework by applying three different eigenvalue tuning strategies. The effects of eigenvalue on influencing the evolution are investigated through combining maximum likelihood estimates of Gaussian model with each of the eigenvalue tuning strategies in ED-EDA. In our experiments, proper eigenvalue tunings show high efficiency in solving problems with small population sizes, which are difficult for classical MGEDA adopting maximum likelihood estimates alone. Previously developed covariance matrix repairing (CMR) methods focusing on repairing computational errors of covariance matrix can be seen as a special eigenvalue tuning strategy. By using the ED-EDA framework, the computational time of CMR methods can be reduced from cubic to linear. Two new efficient CMR methods are proposed. Through explicitly tuning eigenvalues, ED-EDA provides a new approach to develop more efficient Gaussian based EDAs. (c) 2008 Elsevier Inc. All rights reserved.
引用
收藏
页码:3000 / 3023
页数:24
相关论文
共 25 条
[1]  
[Anonymous], 2000, P 2000 GEN EV COMP C
[2]  
[Anonymous], 2004, RR5190 INRIA
[3]  
[Anonymous], 1986, Principle Component Analysis
[4]  
[Anonymous], 1986, Non-Uniform Random Variate Generation
[5]  
[Anonymous], 2000, P GENETIC EVOLUTIONA
[6]  
Bosman P. A. N., 2000, Parallel Problem Solving from Nature PPSN VI. 6th International Conference. Proceedings (Lecture Notes in Computer Science Vol.1917), P767
[7]  
BOSMAN PAN, 2005, MANNH BUS SCH DEP LO
[8]  
BOSMAN PAN, 1999, UUCS199946, P46
[9]  
Bosman PAN, 2007, GECCO 2007: GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, VOL 1 AND 2, P492
[10]  
Chatfield C., 1980, INTRO MULTIVARIATE A