24.12.2014 Views

Earthquake Engineering Research - HKU Libraries - The University ...

Earthquake Engineering Research - HKU Libraries - The University ...

Earthquake Engineering Research - HKU Libraries - The University ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Proceedings of the International Conference on «««<br />

Advances and New Challenges in <strong>Earthquake</strong><br />

<strong>Engineering</strong> <strong>Research</strong>, Hong Kong Volume<br />

A HYBRID OPTIMIZATION ALGORITHM:<br />

GENETIC ALGORITHM - SIMPLEX<br />

HAN Wei 1 ' 2 and LIAO Zhenpeng 1<br />

'institute of <strong>Engineering</strong> Mechanics, China Seismological Bureau,<br />

Harbin, China<br />

"Harbin Institute of Technology, Harbin, China<br />

ABSTRACT<br />

Genetic algorithm is a global optimization algorithm that attracts much attention, and has been used in<br />

many fields recent years. Although it can search widely in model space, its local searching ability is<br />

poor, especially for some complex optimization problems in earthquake engineering. In general, the<br />

local optimization algorithm is excellent in searching ability. To cope with the problem of low<br />

searching efficiency and premature of genetic algorithm, a hybrid global - local optimization<br />

algorithm, genetic algorithm-simplex, is proposed in this paper. Taking the advantages of effective<br />

searching, easy implementing and proper combining of simplex, the proposed algorithm can overcome<br />

the premature of genetic algorithm, and the calculating efficiency of genetic algorithm is also<br />

improved by using simplex to enhance the searching ability, providing new models and increasing the<br />

variety of the population. Numerical experiments were carried out by seven test functions that can<br />

rigorously examine the searching ability of algorithm in various aspects. Compared with several other<br />

global optimization algorithms, the searching ability of the proposed genetic-simplex is verified.<br />

1. INTRODUCTION<br />

Genetic algorithm (GA) is a global optimization algorithm which uses simple encoding technology to<br />

express complex parameters. By genetic operation and natural selection, it can search model space<br />

intelligently. Because of its population searching mode, genetic algorithm can search many locations in<br />

the solution space synchronously. Moreover, natural selection and simple genetic operation make<br />

genetic algorithm easy to use and is not restricted by some additional conditions and information. <strong>The</strong><br />

widely used genetic algorithm is simple genetic algorithm (SGA) with elitist strategy. It employs<br />

binary encoding and roulette wheel selection, and uses just two genetic operations: crossover and

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!