遗传算法与神经网络相结合的降水预报模型  

Precipitation Forecasting Model Based on Neural Network and Genetic Algorithms

作　　者：农吉夫[1,2] 

机构地区：[1]广西民族大学理学院,南宁530006 [2]广西混杂计算与集成电路设计分析重点试验室,南宁530006

出　　处：《大学数学》2012年第5期114-118,共5页College Mathematics

基　　金：国家自然科学基金项目(11061005);广西教育厅科研项目(201204LX083)

摘　　要：针对目前BP神经网络在实际应用中,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,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.同时通过实例说明该模型在降水预测中的应用,计算结果表明该方法的预测精度较高.For overcoming difficulties in application of the method of BP neural network, this paper proposed to optimize the neural network structure and connection weights by means of genetic algorithm whilst to reserve the best individual in evolution process, so that to build up a genetic algorithms Neural Networks model. Through an example we explain the application of precipitation forecasting. The result of study indicates that this method can gain a higher forecasting precision.

关 键 词：短期气候预测 神经网络 遗传算法 

分 类 号：P457.6[天文地球—大气科学及气象学]