摘要
针对目前BP神经网络在实际应用中,网络结构难以确定以及网络极易陷入局部解问题,用遗传算法优化神经网络的连接权和网络结构,并在遗传进化过程中采取保留最佳个体的方法,建立基于遗传算法的BP网络模型,同时用一个实例说明该模型在预测油管钢腐蚀速率中的应用,实践证明神经网络预测结果与实验值吻合较好.最后通过现场实验数据检验了该神经网络的泛化能力,表明其预测结果与现场实验结果相近.
For overcoming difficulties in application of the method of BP neural network, this paper proposed to optimize the neural network structure and connection weights by means of genetic algorithm whilst to reserve the best individual in evolution process, so that to build up a genetic algorithms Neural Networks model. Through an example we explain the application of this model in predicting the corrosion rate of carbon steel. Evidence shows that the predicted values accord with the values of laboratory tests very well. At last, the applicability of the generalization of the model was identified by use of the data from field tests. It shows that the predicted results closed to that of the field tests.
出处
《腐蚀科学与防护技术》
CAS
CSCD
北大核心
2007年第3期225-228,共4页
Corrosion Science and Protection Technology
关键词
神经网络
腐蚀
遗传算法
neural network
corrosion
genetic algorithms
