遗传算法BP神经网络的预报研究和应用

Forecast Research and Applying of BP Neural NetworkBased on Genetic Algorithms

针对目前 BP神经网络在实际气象预报应用中 ,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 ,并以广西的月降水量进行实例分析 ,计算结果表明 ,该方法预报精度高、而且稳定 .

Aim at difficultly deciding the BP network and easily falling into local solution in Meteorological applying, Optimized the neural network and connection weights by means of genetic algorithm and reserved the best individual in evolution process. This method be build up the prediction model. The applied example is monthly mean rainfall the north area of Guangxi in April. Evidence shows the way have high accuracy and stability.