华中科技大学学报（自然科学版）2026，Vol.54Issue(7)：20-26,7.DOI:10.13245/j.hust.250125
动态权值自适应分配的LSTM洪水演进方法
Dynamic weight adaptive allocation-based LSTM method for flood routing
摘要
Abstract
To enhance the simulation accuracy of flood routing and improve the model's capability to capture flood peaks and adjacent flow variations,a dynamic weight adaptive allocation-based long short-term memory(LSTM)method for flood routing was proposed.First,a dynamic weight allocation matrix was established based on the regularization of prediction error values for flood sub-sequences.Subsequently,the dynamic weights were incorporated into the mean squared error(MSE)loss function to formulate a model optimization strategy with adaptive learning weights.Such strategy enables the model to adaptively amplify learning weights for high-flow periods while attenuating those for low-flow periods.To validate the proposed method,a case study was carried out at the lower reaches of the Hanjiang River from Huangzhuang to Xiantao hydrologic station.Results demonstrated that,compared to typical LSTM approaches,the proposed method significantly improved simulation accuracy for flood peaks and high-flow periods while exhibiting more stable performance across the entire flood process,markedly enhancing the overall effectiveness of flood routing.关键词
洪水演进/深度学习/长短期记忆网络(LSTM)/动态权值/自适应分配Key words
flood routing/deep learning/long short-term memory(LSTM)/dynamic weight/adaptive allocation分类
建筑与水利引用本文复制引用
姜维龙,陈璐,易彬,康乐..动态权值自适应分配的LSTM洪水演进方法[J].华中科技大学学报（自然科学版）,2026,54(7):20-26,7.基金项目
西藏自治区自然科学基金资助项目(XZ202401ZR0044) （XZ202401ZR0044）
国家自然科学基金联合基金资助项目(U24B20105) （U24B20105）
西藏农牧学院人才队伍建设项目(XZNMXYRCDWJS-2024-001) （XZNMXYRCDWJS-2024-001）
国家资助博士后研究人员计划(GZC20252078). （GZC20252078）
