BP-GA学习算法在短期洪水预报中的应用  

Application of BP-GA mixed algorithms to forecasting short term flood

作　　者：张晔[1] 钱镜林[1] 

机构地区：[1]浙江省水利河口研究院,浙江杭州310020

出　　处：《浙江水利科技》2006年第1期25-28,共4页Zhejiang Hydrotechnics

摘　　要：先用遗传算法在权值空间中搜索,直到找到某一稳定的适应度值。针对遗传算法(GA)只能找到问题的本致解的缺点,利用BP算法在GA找到的权空间中进一步搜索,以求找至问题的准确解,并实现短期降雨径流的模拟和预报。在构建网络模型时,综合考虑了气候因素、前期降雨量上游支流的前期流入流量和计算出口断面的前期流量的影响。实验结果表明,基于这一方法的短期洪水预报具有较高的精度和可靠性。BP networks has been apphed in rainfall-runoff simulation widely. But it is right correlative with the initial state whether the network is convergent or not. To avoid getting into the local minimum, the GA was applied to get the initial weights in this paper firstly till the leaming error has tended to stability, Here , the best initial weights have been found. Then the BP networks was applied to finish short-term flood forecast process. We also consider the influence of climate, prophase rain, prophase runoff of upriver stream and outlet. Experimental results show that the method based on BP-GA has high precision and is rehable.

关 键 词：BP网络 遗传算法 短期洪水预报 应用 

分 类 号：P333.1[天文地球—水文科学]