群居蜘蛛优化算法在水文频率曲线参数优化中的应用

Application of social spider optimization algorithm in parameter optimization of hydrological frequency curve

利用一种新型群体智能仿生算法——群居蜘蛛优化算法(SSO)优化水文频率曲线参数,以云南省丽江仁里站和总管田站年径流量数据为例进行实例研究,分别将离差平方和准则(OLS)、离差绝对值和准则(ABS)以及相对离差平方和准则(WLS)作为SSO算法最优适应度函数对皮尔逊Ⅲ型分布参数进行优化,优化结果与粒子群优化算法(PSO)、矩法进行对比。结果表明:利用SSO算法优化仁里站和总管田站得到的OLS、ABS、WLS均优于PSO算法及矩法,比矩法提高了11%以上。SSO算法具有收敛速度快、全局寻优能力强等特点,基于SSO算法的优化适线法能够降低水文频率的分析误差,有效提高理论频率曲线与实测数据的拟合精度,是一种可行的水文频率分析方法。

The paper used a new kind of swarm intelligent bionic algorithm---communal spiders optimi-zation ( SSO) to optimize hydrologic frequency curve parameter .Taking the runoff data at Renli and Zong-guantian station along Lijiang of Yunnan Province as an example for case study , the paper respectively let the square error criterion (OLS), deviation of absolute value and criterion (ABS) and relative deviation square and criterion ( WLS) as SSO algorithm optimal adaptation degree function to optimize Pearson type III distribution parameters .Then it compared the optimized results with that of particle group optimization algorithm ( PSO) and moment method .The results show that SSO , WLS and OLS at Renli station and Zongguan station are better than PSO algorithm and ABS algorithm , the result increased by more than 11%.SSO algorithm has fast convergence speed and strong global search optimization ability and so on . Based on that the optimal curve fitting method of SSO algorithm can decrease the analysis error of hydro -logical frequency and effectively improve the accuracy of theoretical frequency curves with measured data fitting ,so it is a feasible h analysis method of ydrological frequency .