基于粒子群神经网络的凌汛开河日期预测研究

为了更准确地预测凌汛开河日期，提出用粒子群算法和BP神经网络相结合的粒子群神经网络模型。介绍了模型的设计和算法实现的流程。该模型通过粒子群算法对BP神经网络初始的权值和阈值进行优化，并以黄河内蒙段三湖河口站作为研究实例进行冰凌开河日期预测。结果表明，经粒子群优化后的BP神经网络预测精度比遗传神经网络和单一BP神经网络更高。

Prediction of beginning date of ice jam flood based on particle swarm optimization-neural network model

In order to predict the beginning date of ice jam flood more accurately,the particle swarm optimization-neural network model is proposed by the combination of particle swarm optimization algorithm and BP neural network.The design of the model and flow of the algorithm realization is presented.The initial weights and thresholds of BP network are optimized by using particle swarm optimization algorithm,and taking the Sanhuhekou Station of Yellow River in the Inner Mongolia as example,the beginning date of ice...