摘要
传统启发式算法在水文模型参数估计中通常存在着易早熟和收敛速度慢等缺陷,为提高水文模型的参数优化精度和算法性能,引入混合蛙跳算法(SFLA),提出一种基于SFLA的水文模型参数估计方法,将该方法应用到新安江模型的参数估计中,并与基于遗传算法(GA)的参数估计方法进行实验对比分析。实验结果表明:基于SFLA的参数优化方法在平均优化精度上相比遗传算法提高了2.5%;在固定优化精度时,优化成功率相比遗传算法提高了53.33%。证明了混合蛙跳算法应用于水文模型参数估计时,在收敛精度和收敛速度方面均有明显优势。
Traditional heuristic algorithms usually have defects such as prematurity and slow convergence for parameter estimation in Hydrologic model. In order to improve the accuracy of parameter optimization and performance of optimization algorithm for hydrological model,Shuffled Frog Leaping Algorithm（ SFLA） was introduced to propose a new method of Hydrological modelparameters estimation based on SFLA. Then the method was applied to estimate the parameters of Xin＇anjiang model,and compared with the parameter estimation method based on Genetic algorithm（ GA） through experimental analysis. Experimental results show that the parameter optimization method based on SFLA could improve 2. 5% of the average optimize accuracy and to GA,and 53.33% of success optimize rate under the fixed precision condition than the GA method. It is proved that SFLA could be applied for parameter estimation in Hydrologic model and it has more obvious advantages in both the speed and the accuracy of convergence.
出处
《重庆理工大学学报（自然科学）》
CAS
2016年第3期80-86,共7页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学基金资助项目(61462058)
兰州交通大学青年科学基金资助项目(2013032)
