基于PSO-AGA的水文频率参数优化算法  

Optimization Algorithm of Hydrologic Frequency Parameters Based on PSO-AGA

作　　者：唐颖[1,2] 张永祥[1,2] 王昊[1,2] 刘宇[1,2] 

机构地区：[1]北京工业大学建筑工程学院,北京100124 [2]北京工业大学水质科学与水环境恢复工程北京市重点实验室,北京100124

出　　处：《北京工业大学学报》2016年第6期953-960,共8页Journal of Beijing University of Technology

基　　金：国家科技支撑计划资助项目(2011BAC12B00)

摘　　要：为了寻求水文频率参数最优值,进而得到精度更高的水文特征值,提出一种基于粒子群优化(particle swarm optimization,PSO)算法与自适应遗传算法(adaptive genetic algorithm,AGA)的水文频率参数优化算法.该算法基于离差平方和最小准则、离差绝对值和最小准则及相对离(残)差平方和最小准则,构建水文频率参数优化模型,在粒子群算法中引入自适应遗传算子,将遗传算法的全局搜索能力强与粒子群算法的收敛速度快有效结合,并对交叉、变异概率进行自适应改进,形成一种自适应混合算法,用于对模型求解,得到最佳水文频率统计参数.以北京市气象中心降雨资料为例,将本文算法与其他常规方法比较,结果表明:本文算法获得的参数值在拟合精度和适线效果上要优于常规方法,为水文频率分析领域决策提供了参考依据.To seek the optimal value for hydrological frequency parameters, and then to obtain a higher precision of hydrological characteristics value, an optimization algorithm of hydrologic frequency parameter based on particle swarm optimization （PS0） and adaptive genetic algorithm （AGA） was proposed. Based on the rule of minimum sum of squared residuals, the rule of the least sum of deviation absolute value and the rule of relative deviation minimum sum of squared residuals, the algorithm was constructed which was applied to hydrological frequency parameter optimization model. Adaptive genetic operators in particle swarm optimization algorithm was introduced, by combining the global search ability of genetic algorithm with quicker convergence rate of particle swarm algorithm effectively, and adaptively, and the crossover and mutation probability was improved, thereby a set of adaptive hybrid algorithm was formed, the optimum parameters of the hydrologic frequency was obtained through the model. By using a municipal meteorological center of rainfall data as an example, the algorithm was compared with other conventional methods in this paper. Results show that the fitting precision and fitness effect of the parameter estimation using the algorithm are superior to conventional methods, and the algorithm provides reference for hydrologic frequency analysis field.

关 键 词：水文频率参数 粒子群算法 遗传算法 适线 

分 类 号：TU992.1[建筑科学—市政工程]