基于混沌优化GMDH网络的月降水量预测

Monthly precipitation prediction of GMDH network based on chaos optimum

为提高降水量预测的精确度,介绍了一种基于混沌优化的GMDH网络预测方法,该方法利用混沌优化算法全局搜索GMDH网络的初始权值,并利用优化后的GMDH网络建立预测模型对月降水量进行预测。结果表明:该方法能加快GMDH网络结构稳定的速度,使算法快速收敛到全局最优解,对月降水量的动态预报具有一定的实用价值。

This paper introduced a GMDH network prediction method based on Chaos optimum,in order to increase the accurate of precipitation forecast.It used chaos optimum method to globally search GMDH network＇s initial weight value,and applied the method to predict monthly precipitation.The simulation results showed that the method can accelerrate the stabilization of GMDH network＇s structure,converge to the global optimal solution quickly and has some practical value for dynamic forecast of monthly precipitation.