基于改进遗传算法的BISQ模型多参数反演方法研究  

The methodological study of multi-inversion of BISQ mold parameters based on modified genetic algorithm

作　　者：吴涛[1,2] 李红星[3] 陶春辉[1,2] 顾春华[1,2] 

机构地区：[1]国家海洋局第二海洋研究所,杭州310012 [2]国家海洋局海底科学重点实验室,杭州310012 [3]东华理工大学核工程技术学院,抚州344000

出　　处：《地球物理学进展》2012年第5期2128-2137,共10页Progress in Geophysics

基　　金："973"国家重点基础研究发展计划(2012CB417305);国家自然科学基金项目(41104073);海洋公益性行业科研专项经费项目(200805005);江西省自然科学基金项目(2010GQS0002);国家海洋局海底科学重点实验室开放基金(KCSG0803);国家海洋局青年海洋科学基金资助项目(2011326);国家海洋局第二海洋研究所基本科研业务费专项资助(JG1005)联合资助

摘　　要：BISQ模型反演在油气资源与海洋地震勘探中具有广阔的应用前景与实用价值.为解决具高维参数与多峰值的BISQ模型参数反演问题,本文在传统的遗传算法基础上添加插入了全局保优机制、Boltzmann生存机制、小生境之DC漂移技术和群体灾变等方法与技术,使改进组合的遗传算法能在高维参数与多峰值的反问题中搜索得到全局最优解.结合海底多频原位沉积物声学测量数据,应用本文方法,得到了与声波传播速度相似性最优模型的各参数值,效果明显.BISQ mold inversion has vast vista application and practical value in the exploration of ocean oil and gas resources,and Marine seismic survey fields. Intending to solve the problem of parameters inversion for BISQ mold with high dimensional parameters and multimodal function, based on the traditional genetic algorithm, this paper added and inseted the global best mechanism, the Boltzmann existent mechanism, the DC drifting of niching technology, the idea of catastrophe of population, and other methods and techniques. The modified combined genetic algorithm, used in a high dimensional parameters and inversion problem with multimodal function, can help to get global optimal values. Using the method introduced in this paper, and combining with the in-situ experimental data of multi-frequency acoustic velocities of submarine sediment, have got the parameters values which well compared with the acoustic velocities, and obtained a satisfied results.

关 键 词：改进组合 遗传算法 BISQ 多参数反演 灾变 

分 类 号：P631[天文地球—地质矿产勘探]