优选伸缩比SCE-UA算法与新安江模型参数率定研究  

Study on the SCE-UA algorithm combined with optimal scaling factor and parameter calibration of the Xinanjiang model

作　　者：侯宇 阚光远[1,2,3,4] 梁珂 HOU Yu;KAN Guangyuan;LIANG Ke(China Institute of Water Resources and Hydropower Research,Beijing 100038,China;Research Center on Flood&Drought Disaster Prevention and Reduction of the Ministry of Water Resources,Beijing 100038,China;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,Beijing 100038,China;Key Laboratory of Water Safety for Beijing-Tianjin-Hebei Region of Ministry of Water Resources,Beijing 100038,China;Beijing IWHR Corporation,Beijing 100048,China)

机构地区：[1]中国水利水电科学研究院,北京100038 [2]水利部防洪抗旱减灾工程技术研究中心(水旱灾害防御中心),北京100038 [3]流域水循环模拟与调控国家重点实验室,北京100038 [4]水利部京津冀水安全重点实验室,北京100038 [5]北京中水科工程集团有限公司,北京100048

出　　处：《中国水利水电科学研究院学报（中英文）》2024年第2期210-218,共9页Journal of China Institute of Water Resources and Hydropower Research

基　　金：国家重点研发计划项目(2023YFC3209202);水利部重大科技项目(SKR-2022056);光合基金A类(ghfund202302018283);中国水利水电科学研究院十四五“五大人才”计划(JZ0199A032021);中国水科院减灾中心“基础研究型”科技创新人才项目(GY2205)。

摘　　要：洗牌复合形进化(SCE-UA)算法作为寻优能力较强的全局寻优算法已被广泛应用于水文模型参数率定领域。经典SCE-UA算法中的下山单纯形搜索采用固定伸缩比实现反射和收缩操作,其寻优效率还有待提升,实际应用中发现变动伸缩比能够改善算法的寻优效率。针对以上问题,结合数值实验与分析,发现伸缩比取值存在最优区间,在伸缩比最优区间内均匀采样,利用优选法在单纯形搜索过程中动态确定最适伸缩比,提出了基于优选伸缩比的SCE-UA算法。本研究利用改进的算法,基于多组人工降雨—径流资料开展了新安江模型参数率定研究。结果表明,相较于固定伸缩比,优选伸缩比算法在多组人工降雨—径流资料下的参数率定测试中均表现出相同迭代次数下目标函数值更优的结果,显著提升了优化效率,进一步提升了算法的深度搜索能力,具有推广应用价值。The Shuffled Complex Evolution-developed at University of Arizona(SCE-UA)algorithm has been widely applied in the field of hydrological model parameter calibration.The algorithm exhibits good robustness property and strong global searching capability.The reflection and contraction operations in the downhill simplex search of the traditional SCE-UA algorithm adopt a fixed scaling factor and its searching efficiency has potential to be further improved.In real-world applications,we find that the searching efficiency can be enhanced by adjusting the scaling factor.According to numerical experiment and analysis,it is recognized that there is an optimal interval for the scaling factor.In order to cope with the above-mentioned problems,the improvement of the SCE-UA algorithm by combining optimal scaling factor is proposed,which uniformly samples the scaling factor values in a discrete manner within the optimal interval and dynamically selects the optimal factor value during the process of parameter optimization.Studies on Xinanjiang model parameter calibration are carried out based on the synthetic rainfall-runoff data and the proposed optimization algorithm.The research results indicate that the improved algorithm,which incorporates optimal scaling factor,significantly improves the optimization efficiency compared to the utilization of a fixed value.This improvement holds valuable potential for future general purpose real-world applications.

关 键 词：SCE-UA 下山单纯形搜索 伸缩比 优选法 动态 

分 类 号：P333[天文地球—水文科学]