中文摘要：

采用差分进化自适应Metropolis（DREAM）算法对描述含水层渗透系数空间变异特征的参数进行识别.利用直接傅里叶变换方法产生一组空间结构参数下的若干个渗透系数场实现,借鉴噪声遗传算法（NGA）思想,计算该组空间结构参数对应的贝叶斯后验概率值,提高DREAM算法求解的效率.算例求解结果表明,DREAM算法能够有效获得含水层渗透系数空间结构参数的后验分布,并可得到对应的一系列渗透系数场,可为含水层参数空间变异性研究提供新的思路.

英文摘要：

In this study,it adopted the Differential Evolution Adaptive Metropolis（DREAM）algorithm to identify the spatial structure parameters of the spatial variability of aquifer hydraulic conductivity.The direct Fourier transform method is used to generate a certain number of hydraulic conductivity field realizations under a special set of the spatial structure parameters.Furthermore,the idea of Noisy Genetic Algorithm（NGA）is introduced to calculate the Bayesian posterior probability values,and improve the computational efficiency of the DREAM algorithm.Optimization results of the field application shows that the DREAM algorithm can effectively find the posterior distribution of the spatial structure parameters of the aquifer hydraulic conductivity,and a set of hydraulic conductivity field realizations can be obtained.Therefore,it may provide a novel research idea for identification of the spatial variability of aquifer hydraulic conductivity.