基于GA的RBF网络用于旱涝灾害因素预测研究

DISASTER ELEMENTS PREDICTION WITH RBF NETWORK BASED ON GA

用基于GA的RBF网络构成非线性时间序列预测器 ,针对灾害因素中的降雨量、最高水位、最大流量等进行预测。在运用GA的过程中 ,针对RBF的网络结构提出了与以往不同的编码方式 ,使得整个编码过程简单有效 ,而且符合RBF网络本身的特性。

We use RBF neural network based on GA to make up a nonlinear time series predication implement, which is used to predict the disaster elements, such as the precipitation, the highest water level and the biggest flow, etc. We advance a new code way in accordance with the RBF neural network, which not only made the whole code process simply and efficient but also accord with the characteristic of RBFNN.