摘要
围岩参数优化反演实质为一个高度复杂的非线性函数优化问题,采用传统局部优化技术存在依赖于初值、易失败等缺点,而进化算法这种全局优化方法却可以很好地解决此问题。但常用的遗传算法存在效率低等缺点,结合免疫系统原理及进化规划机制提出的新型仿生算法——免疫进化规划可以克服遗传算法的缺点。通过某水电站工程洞室围岩参数反演研究,证明了仿生算法参数反演的合理性、有效性。
The objective function of the optimization parameter back analysis is a very complicated non-linear multiple-hump function. To solve this problem, the traditional optimization methods have some problems, such as reliance of the result on initial values, proness to fail, while the evolutionary algorithm—a global optimization method can solve it very well. However, the efficiency of the generally used genetic algorithm (GA) is very low. To overcome this difficulty, a new bionics algorithm: immune evolutionary programming (IEP) which combines artificial immune system (AIS) and evolutionary programming (EP) is proposed. At last, this new parameter inversion method based on IEP is applied to an engineering of a power station.
出处
《岩石力学与工程学报》
EI
CAS
CSCD
北大核心
2002年第A02期2521-2526,共6页
Chinese Journal of Rock Mechanics and Engineering
关键词
地下工程
参数反演
优化方法
进化算法
仿生算法
underground engineering
parameter inversion
optimization method
evolutionary algorithm
bionics algorithm
