摘要
介绍了BP神经网络应用于损伤诊断时的局限性,将全局寻优能力很强的遗传算法作为其改进算法来优化神经网络的初始权值和阈值,并将改进的遗传神经网络应用于桁架结构的损伤诊断,数值模拟试验证实了方法的有效性。
The setbacks of the back propagation algorithm is discussed and the genetic algorithm is presented to improve the ability of the back propagation neural networks, then the improved algorithm is used for damage detection of a truss system, the result of the numerical simulation is solid and shows the effectiveness of the improved algorithm.
出处
《建筑技术开发》
2008年第6期6-7,24,共3页
Building Technology Development
关键词
遗传神经网络
遗传算法
BP神经网络
损伤诊断
Genetic artificial neural network
Genetic algorithm
BP neural network
Damage detection
