河道水量还原计算中的遗传BP人工神经网络方法  

River water returning calculation based on genetic BP neural network

作　　者：魏茹生[1] 沈冰[1] 吴建华[2] 

机构地区：[1]西安理工大学西北水资源与环境生态教育部重点实验室,陕西西安710048 [2]太原理工大学,山西太原030024

出　　处：《河海大学学报（自然科学版）》2008年第2期161-164,共4页Journal of Hohai University(Natural Sciences)

摘　　要：将遗传算法GA及人工神经网络BP算法结合起来运用于河道水量的还原计算.结果表明:基于GA+BP的河道水量还原优化算法能够克服BP算法自身不可优化的弊病,较好地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.同时该算法较确切地反映出河道水量还原计算中河道水量形成非线性这一本质现象,提高了河道水量还原计算结果的精度.Returning calculation of river water was carried out on the basis of GA （genetic algorithm） and BP （back propagation） algorithm of artificial neural networks. Results indicate that the optimization algorithm based on GA ＋ BP, which does not have the defect that BP algorithm cannot be optimized, can improve the global optimization ability of the net, increase the net speed, and avoid the local minimum value in the net. The algorithm reflects the nonlinear character of water formation in river water returning calculation, and improves the calculation precision.

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]