Accelerating evolutionary algorithms with Gaussian process fitness function models

被引:247
作者
Büche, D [1 ]
Schraudolph, NN [1 ]
Koumoutsakos, P [1 ]
机构
[1] Swiss Fed Inst Technol, Inst Computat Sci, Zurich, Switzerland
来源
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART C-APPLICATIONS AND REVIEWS | 2005年 / 35卷 / 02期
关键词
evolution control; evolutionary algorithms (EAs); fitness function modeling; gas turbine compressor design; Gaussian process; surrogate approach;
D O I
10.1109/TSMCC.2004.841917
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We present an overview of evolutionary algorithms that use empirical models of the fitness function to accelerate convergence, distinguishing between evolution control and the surrogate approach. We describe the Gaussian process model and propose using it as an inexpensive fitness function surrogate. Implementation issues such as efficient and numerically stable computation, exploration versus exploitation, local modeling, multiple objectives and constraints, and failed evaluations are addressed. Our resulting Gaussian process optimization procedure clearly outperforms other evolutionary strategies on standard test functions as well as on a real-world problem: the optimization of stationary gas turbine compressor profiles.
引用
收藏
页码:183 / 194
页数:12
相关论文
共 36 条
[1]  
[Anonymous], MANUAL TPROS V2 0
[2]  
[Anonymous], 9833 ICASE NASA LANG
[3]  
[Anonymous], P 7 AIAA USAF NASA I
[4]   Gaussian process modelling of austenite formation in steel [J].
Bailer-Jones, CAL ;
Bhadeshia, HKDH ;
MacKay, DJC .
MATERIALS SCIENCE AND TECHNOLOGY, 1999, 15 (03) :287-294
[5]  
Barnett S., 1979, Matrix methods for engineers and scientists
[6]   A rigorous framework for optimization of expensive functions by surrogates [J].
Booker A.J. ;
Dennis Jr. J.E. ;
Frank P.D. ;
Serafini D.B. ;
Torczon V. ;
Trosset M.W. .
Structural optimization, 1999, 17 (1) :1-13
[7]  
BUCHE D, 2004, THESIS ETH ZURICH SW
[8]  
BUCHE D, 2003, P ASME IGTI TURB EXP
[9]  
DENNIS J, 1993, COMPUTATIONAL MATH
[10]  
DRELA M, 1995, USERS GUIDE MISES 2