地球物理反演中的差分进化算法

为避免传统线性反演过程中极度依赖初始模型的选取、易陷入局部最优解的问题，在综合研究现有地球物理非线性反演方法的基础上，引入了非线性领域群体智能算法—差分进化算法，并以电测深方法为例，建立起了差分进化算法的非线性反演模型，通过对所建模型试算及实测数据验证，证实该方法具有反演速度快、精度高等优点，并可推广应用于二维或三维电测深反演理论之中。最后将该算法在内蒙古某多金属矿勘查中进行了实际推广，取得了较好的效果。

Applying the differential evolution algorithm to geophysical inversion

In this study, we introduced the differential evolution (DE) algorithm in the geophysical nonlinear inversion field to avoid the deficiency of the traditional linear inversion methods. We have established a nonlinear inversion model of DC electrical sounding data using the DE algorithm. A comparison between the model and actual data suggests that the DE algorithm is more accurate and faster than existing algorithms. The DE algorithm can be further applied to two-dimensional or three-dimensional electrical sounding inversion. We also applied the DE algorithm to the polymetallic ore exploration successfully in Inner Mongolia.