人工蜂群算法的电震非线性联合反演研究  

Application of artificial bee colony algorithm in electrical-seismic nonlinear joint inversion

作　　者：侯征[1,2] 熊盛青[1] 杨进[3] 于长春[1] 刘国辉[2] 

机构地区：[1]中国国土资源航空物探遥感中心,北京100083 [2]石家庄经济学院勘查技术与工程学院,石家庄050031 [3]中国地质大学,北京100083

出　　处：《地球物理学进展》2015年第6期2666-2675,共10页Progress in Geophysics

基　　金：国家高技术研究发展计划(863)(2013AA0639)之"航空地球物理遥感综合探测技术与解释系统研究"(2013AA063905);国家自然科学基金项目(41130637);国家自然科学基金项目(41374133);河北省科技支撑计划项目(13274216)联合资助

摘　　要：本文为解决直流电阻率法和瑞雷面波法反演存在的多解性问题,构建了新的电震联合反演目标函数,通过增加不同地球物理响应的反演参数,利用权系数调节各参数对反演结果的影响,实现同步联合反演,从而减少多解性.引入人工蜂群算法求解目标函数,有效解决了传统反演方法将非线性问题线性化处理产生的细节信息丢失问题,提高了反演精度.经过对较为复杂的四层速度递增AA型和四层含低速夹层HA型理论模型和实例数据的反演验证,表明基于人工蜂群算法的非线性联合反演能够有效减少单一方法多解性问题并提高反演精度.In this paper,we establish a new joint inversion objective function to solve the multiple solutions problem in DC resistivity method and Rayleigh wave method,by adding different geophysical response inversion parameters,using weighting coefficient to adjust the influence of various parameters on the inversion results to realize the synchronous joint inversion.We introduce artificial bee colony algorithm to solve the objective function.Artificial bee colony algorithm effectively solves details information loss problem producing in linearization of nonlinear problems.This method improves the inversion accuracy.After inversion validation for four layers velocity increment AA model, four layers low velocity interlayer HA model and instance data,shows that nonlinear joint inversion based on artificial bee colony algorithm can effectively reduce multiple solutions of single method and improve the inversion accuracy.

关 键 词：非线性联合反演 直流电阻率测深法 瑞雷面波法 人工蜂群算法 

分 类 号：P631[天文地球—地质矿产勘探]