基于GA+BP的河道水量还原的优化算法研究  

Research on the Optimization Algorithm of River Water Reduction Based on GA+BP

作　　者：刘京铄[1] 李光中[1] 

机构地区：[1]湖南工程学院应用技术学院,湖南湘潭411101

出　　处：《中国农村水利水电》2009年第5期26-28,33,共4页China Rural Water and Hydropower

基　　金：山西省自然科学基金(2006011059)

摘　　要：在充分理解河道水量还原、GA及BP算法的基础上,根据水量还原计算中水文现象的本质规律,将三者相结合,并运用于实际。检验证明:基于GA+BP的河道水量还原的优化算法能够克服BP算法自身不可优化的弊病,大大改善网络全局寻优能力,提高网络速度,防止网络陷入局部最小值。同时该算法更加真实的映射出河道水量还原计算中水文现象的本质规律,提高了河道水量还原计算结果的精度。On the basis of understanding stream channel water reduction, GA（genetic algorithm）, BP（back propagation of artificial neural networks） and the principle rule of hydrograph phenomenon in reduction calculation, the three points are combined and applied in practice.. Practice shows that the majorized algorithm based on GA＋BP can overcome the shortcoming that BP algorithm can＇t be optimized. It can also improve the capacity of seeking the best result in the whole net, increase net speed, and prevent the net trap in the partial minimization. Such majorized algorithm can reflect the innate character of hydrograph phenomenon in river water reduction calculation, improve the accuracy of reduction calculation.

关 键 词：人工神经网络 还原计算 遗传算法(GA) BP模型 

分 类 号：P333.1[天文地球—水文科学]