用差异演化-粒子群混合算法确定含水层参数

(长安大学 理学院,陕西 西安 710064)

含水层参数; 差异演化算法; 粒子群优化算法; 直线供水边界

Differential evolution-particle swarm optimization mixed algorithm determine aquifer parameters
DUAN Guo-rong,LIU Yuan-hui

(College of Science,Chang'an University,Xi'an 710064,China)

aquifer parameters; differential evolution algorithm; particle swarm optimization algorithm; linear pervious boundary

DOI: 10.13800/j.cnki.xakjdxxb.2019.0323

备注

差异演化-粒子群混合优化算法是利用差异演化算法变异粒子个体的历史最优位置,保持种群多样性,与粒子群优化算法相结合的一种混合优化算法。应用差异演化-粒子群混合优化算法分析直线供水边界含水层抽水试验数据确定含水层参数,并将该算法所得计算结果与其他方法的结果进行比较,分析待估参数的不同取值范围对参数估计的影响,发现差异演化-粒子群混合优化算法能够有效地确定含水层参数,且目标函数值更小,计算精度更高,可达到1.673 1*10-6; 待估参数范围的增大对差异演化-粒子群混合优化算法的收敛性影响较小,当待估参数取值范围的上限扩大到原来的14倍时,收敛率高达97%,且得到目标函数值基本一致,由此可知,差异演化-粒子群混合优化算法对初值选取的敏感性低,寻优能力强,稳定性好。数值实验结果表明:差异演化-粒子群混合优化算法有效地避免了粒子群优化算法所出现的早熟现象,是分析抽水试验数据,确定含水层参数和计算观测孔与虚拟映射井之间距离的有效方法。

The differential evolution-particle swarm optimization mixed algorithm is a hybrid optimization algorithmcombiningwith particle swarm optimization algorithm andusing difference evolution algorithm to mutate individual historical best position to keep population diversity.It is applied in the analysis of pumping experiment data in linear pervious boundary to estimate the aquifer parameters.The calculation results are compared with the results of other methods,and influence of the range of initial guess value on estimating parameter is analyzed.It is pointed out that the differential evolution-particle swarm optimization mixed algorithm can effectively determine aquifer parameters with thesmallertarget function value and the higher calculation precision,which can reach 1.673 1*10-6.The wide range of initial values has less effect on the convergence.When the limit of the value range of the estimated parameter is increased to 14 times,the convergence rate is as high as 97%,and the target function value is basically the same.Thus the algorithm has lower sensitivity to initial selection,and better ability of searching and good stability.The results indicate it can effectively avoid the premature phenomenon of particle swarm optimization algorithm.Therefore the differential evolution-particle swarm optimization mixed algorithm is an effective method to analyze the pumping experiment data for determination of the aquifer parameters and calculation of the distance between the observation hole and the virtual mapping well.