流域水文模型参数识别的现代优化方法研究进展  

Modern Optimization Method of Parameter Identify in Hydrological Models:Research and Application

作　　者：武震[1] 张世强[1] 张小文[2] 

机构地区：[1]中国科学院寒区旱区环境与工程研究所,甘肃兰州730000 [2]兰州商学院,甘肃兰州730020

出　　处：《冰川冻土》2008年第1期64-71,共8页Journal of Glaciology and Geocryology

基　　金：国家重点基础研究发展计划(973计划)资助项目(2007CB411502);国家自然科学基金项目(40671041);中国科学院寒区旱区环境与工程研究所人才基金项目(2004116)资助

摘　　要：水文参数优化是水文模型发展中的一个重要部分,针对水循环过程模拟中的流域水文模型参数识别的复杂优化问题,回顾了水文模型优化算法的发展历程,对国内外水文模型参数识别算法的最新进展进行了阐述.介绍了模拟退火算法、遗传算法、SCE-UA、粒子群算法等常用算法的参数设定及一般流程,并对编码遗传算法、单纯形混合加速遗传算法和运用融合技术的几种遗传算法以及其它几种常用算法在新安江模型中的优化结果进行了比较,认为混合加速遗传算法是一种较好的方法.在系统研究现代优化算法与传统优化算法的基础上,建立各种优化算法的融合技术和法则可能是进一步提高参数优化算法性能的方向.Parameter optimization in hydrology is an important aspect of developing hydrological models. In this paper, on the base of complex optimize problems with regard to runoff hydrological model parameter identify in water cycle simulation, the development process of optimal algorithm with hydrological models is reviewed. Particularly, the new advance in calculation of hydrological model parameter identification at home and abroad is introduced, such as the simulated annealing algorithm, genetic algorithm, SCE-UA algorithm, SCPSO algorithm and coding based accelerating genetic algorithm, simplex algorithm. The optimal results of several genetic algorithms using amalgamation technology and other several algorithms, which are used usually in Xin＇an River Model, are also compared. It is found that the hybrid accelerating genetic algorithm is the predominant. To build various amalgamation technologies and laws is a way, which, based on systemic research modern optimal method and traditional optimal method, may enhance the capability of parameter optimization algorithms.

关 键 词：水文模型 参数识别 算法 

分 类 号：P333[天文地球—水文科学]