非线性位移时间序列进化神经网络建模的适应性研究

Study on Applicability of Genetic-Neural Network Modeling of Nonlinear Displacement Time Series

探讨了非线性位移时间序列建模的进化神经网络方法及其可用性，主要包括进化参数如群体规模、突变变异率和蠕变变异率、随机数种子和个体的选择策略对进化神经网络模型性能的影响。结果表明，对于给定的问题，只要遗传参数取值合理，就可以获得理想的效果。

The application of evolutionary neural network to nonlinear time series ahalsis of slope displacement is studied, and the the of genetic parameters such as population size ,elitism strategy ,creep mutation rate, and jump mutation rate, etc. on the performance of evolutioary neural network model is also tested.The results indicate that desirable results can be obtained for a specific problem if the geetic parameter has a reasonable value .