波阻抗混合反演全局寻优与编码的实验研究

针对遗传算法用于波阻抗反演涉及多参数编码、以及对应遗传操作不易实现的问题,在广泛研究了遗传算法的各种编码方式、及其对应的选择方式、交叉方式以及变异方式特点的基础上,提出了采用二进制字符串编码、遗传操作为最优保存策略选择、多点交叉和非均匀变异的遗传算法,基于该算法形成了全局寻优的波阻抗混合反演方法。文中介绍了遗传算法的研究现状,例举了全局寻优交叉和变异的计算结果,给出了二进制字符串编码的主要程序段,用数值模型与实际资料验正了混合反演算法的有效性。     

量子遗传算法在大地电磁反演中的应用

量子遗传算法（QGA）以量子理论为基础,通过利用量子位编码代替经典遗传算法的二进制位编码,利用量子旋转门定向更新种群来代替传统方法中种群的选择、交叉和变异过程,使得算法具有一定的内在并行运算能力和量子的隧道效应,从而加快了搜索速度,改善了收敛速度,并具有更强的全局寻优能力.本文针对地球物理反演问题的非线性、多极值特点提出一套实现方案,通过理论模型和实测数据试验对比研究,表明量子遗传方法在大地电磁反演中的寻优质量和效果明显优于传统遗传算法.     

基于子空间的二维大地电磁量子遗传反演法研究

量子遗传算法作为一种高效的优化算法,仍存在容易陷入局部极值的缺点.为提高算法的高效性,并探讨将算法应用于大地电磁二维反演的可行性和有效性,本文对算法进行了改进,并通过一维两层D型和四层HK型模型数值试验验证了改进的有效性.然后将改进后的算法引入二维大地电磁反演,在引入滑动子空间思想,同时只考虑最简化反演条件的前提下,对...     

多导体柱微波成像的连续编码遗传算法

提出一种能同时重建多个二维良导体目标外形轮廓的新方法——连续编码遗传算法．将目标的横截面轮廓近似表达为三角级数形式,由边界条件得到一积分方程组,在此基础上将成像问题转化为约束优化问题,级数的各项系数为待优化量．积分方程组为约束条件,目标函数定义为实际测量的散射场与反演过程中得到的散射场之间的相对误差函数．采用连续编码遗传算法求解,优化过程通过选择、交叉、变异等遗传操作的选代而实现,待优化目标函数进行线性变换并采用模拟退火原理确定目标后表达为适应度函数,采用联赛选择与比例选择相结合的选择机制和单点交叉方式,变异操作通过对基因施加微小随机扰动实现．上一代中适应度最高的个体直接保留．数值模拟反演实验验证了方法的有效性．与其他反演方法相比,本法具有简单、通用、鲁棒性强等特点．     

结构模糊控制规则优化生成的遗传算法

本文采用实数个体编码解码、两点交叉、两点变异、保留最优个体的模糊遗传算法对模糊控制规则进行优化;其次,对三层框架结构的模糊遗传算法控制进行了仿真实验,同时,与经验规则的模糊控制效果进行了比较;最后,对结构模糊遗传算法的鲁棒性进行了仿真试验,进一步验证了本文所提方法的可行性和有效性.     

实数编码遗传算法在大地电磁测深二维反演中的应用

作为全局非线性优化的新方法之一的遗传算法,近年来已从生物工程流行到大地电磁测深资料解释中.然而,大地电磁反演问题具有不适定性,解的非唯一性.通过结合求解不适定问题的Tikhonov正则化方法,本文采用实数编码遗传算法求解大地电磁二维反演问题.此算法在构建目标函数时引入正则化的思想,利用遗传算法求解最优化问题.常规的基于局部线性化的最优化反演方法易使解陷入局部极小值,而且严重的依赖初始模型的选择.与传统线性化的迭代反演方法相比,实数编码遗传算法能够克服传统方法的不足且能获得更好的反演结果.通过对大地电磁测深理论模型进行计算,结果表明:该算法具有收敛速度快、解的精度高和避免出现早熟等优点,可用于大地电磁资料解释.     

自适应遗传算法配准三维数字样品模型及其CAD模型

通过包围盒的初始匹配和浮点数编码自适应遗传算法的精确匹配,实现了三维数字样品模型及其CAD模型的配准。初始匹配决定精确匹配空间变换参数的范围;精确匹配在搜索变换矩阵的6个变换参数过程中,采用最小二乘方法构造目标函数;遗传算法中引入种群多样性指标;在遗传操作的交叉算子和变异算子中引入该指标,并将该指标用于交叉概率和变异概率两个参数的调整,有效地克服了遗传算法的早收敛问题。实验结果表明该方法较好地解决了三维数字样品模型及其CAD模型的配准问题,且算法稳定,易于实现。     

多种物理机制耦合作用下的储层介质参数反演研究

孔隙介质的黏弹性、孔隙流体的Biot流动和喷射流动是影响波传播的重要物理机制.本文分别基于弹性和黏弹性BISQ模型,利用自适应杂交遗传算法研究了多种物理机制耦合作用条件下储层介质参数反演.为了测试自适应杂交遗传算法的有效性,本文分别利用自适应杂交遗传算法和传统实数编码遗传算法对含有不同噪声的理论合成数据进行了反演试算.对比理论合成数据反演结果可知,自适应杂交遗传算法具有抗干扰能力强且收敛速度快的特点,是一种有效的储层介质参数反演方法.同时本文也利用不同频率尺度和不同温度条件下的P波和S波实测数据进行了联合反演.对比研究表明,黏弹性BISQ模型能够很好地解释不同频率尺度的波频散特征,不仅能够很好地预测P波速度,而且也能够很好地预测S波速度,从而证明了黏弹性BISQ模型能够准确地描述低频条件下的波频散.     

基于交叉梯度约束的重力、磁法和大地电磁三维联合反演

为了降低单一地球物理方法反演的多解性及受噪声的影响程度,本文围绕重力、磁法和大地电磁法开展了三维联合反演的研究.重、磁采用基于对数障碍法的正则化反演算法,大地电磁使用limited-memory BroydenFletcher-Goldfarb-Shanno(L-BFGS)反演算法,引入交叉梯度函数实现了三种物性结构的相互耦合,最终开发出一套重磁电三维联合反演算法,并实现MPI并行加速计算.通过理论模型算例验证了算法的准确性,结果表明:不论是单棱柱体模型还是组合棱柱体模型,联合反演结果相较单独反演对于异常体的空间形态刻画以及物性数值恢复具有较好的提升;单棱柱体模型算例使得异常体的物性参数(密度、磁化率和电阻率)更加接近于真实的物性参数;组合棱柱体模型的联合反演结果不仅仅消除了围岩物性参数的假异常,而且还增强了异常体边界结构的恢复程度.     

一种波速结构的两步优化反演策略

借助虚拟反演思路,通过对各种遗传算子不同匹配方式的比较研究,指出了对于频散曲线反演浮点数编码与轮盘赌选择的匹配方式离线性能最好,提出了一种两步优化反演策略.该两步策略利用浮点数编码、轮盘赌选择、浮点数均匀交换与变异算子匹配组成基本遗传算法框架,在此框架基础上施加免疫启发策略和免重复计算加速策略,多次运行,对每次运行结果继续施加模拟退火算法使其至少达到局部最优,最后取得最优解.免疫启发策略充分利用最佳个体的信息加速进化进程,通过对每代的最佳个体施加一服从标准正态分布的随机数来加强对邻近区域的局部搜索,通过标准差的调整也兼顾了对邻近区域以外区域的搜索,将局部搜索和全局搜索有机地结合起来,同时还最大限度地降低了对遗传算法自身进化进程的干扰;免重复计算策略大大减少了正演计算次数,节约了计算成本,提高了反演效率.两步反演策略避免了多次平均法的缺陷,提高了反演结果的稳定性和精度,降低了非惟一性.     

基于遗传算法CSAMT反演计算研究

提出了CSAMT方法测深的遗传算法反演,比较传统的线性反演方法,其对初始模型依赖性小,全局寻优能力强,适用于在地下电性参数比较模糊的条件下对测深曲线进行反演.本文设计多种理论模型进行模拟,证明该算法精度高,可行性强.     

一种改进的全局优化算法及其在面波频散曲线反演中的应用

建议一种SA和GA相结合的策略,较好地解决了GA收敛早熟及SA搜索效率较低的问题,提高了全局优化计算效率;在应用其依据面波频散曲线反演工程场地剪切波速时,利用简化剥层法提供较小的模型空间,取得了较好的反演效果.     

Studies on structural vibration control with MR dampers using μGA

A new approach to reducing the seismic response of spatial structures with magneto-rheological (MR) dampers is presented in this paper. The Genetic Algorithm with small populations (μGA) is used to optimize the control for the MR dampers to reduce structural vibration, which is difficult to achieve using classical optimal control. The advantages of μGA are the use of global properties and that fewer conditions are required to obtain the optimal function. Numerical results demonstrate the effectiveness of the proposed method in reducing the seismic response of structures.     

关于-遗传算法收敛性的注记

遗传算法是一种受到广泛注意的全局优化算法,已经在包括地震工程的很多领域中获得应用.本文将结合这一算法的实际操作步骤,对简单遗传算法的不收敛性和包含最优个体保护策略的遗传算法的收敛性给出一个简要的证明.     

Multi-tier interactive genetic algorithms for the optimization of long-term reservoir operation

Genetic algorithms (GAs) are well known optimization methods. However, complicated systems with high dimensional variables, such as long-term reservoir operation, usually prevent the methods from reaching optimal solutions. This study proposes a multi-tier interactive genetic algorithm (MIGA) which decomposes a complicated system (long series) into several small-scale sub-systems (sub-series) with GA applied to each sub-system and the multi-tier (key) information mutually interacts among individual sub-systems to find the optimal solution of long-term reservoir operation. To retain the integrity of the original system, over the multi-tier architecture, an operation strategy is designed to concatenate the primary tier and the allocation tiers by providing key information from the primary tier to the allocation tiers when initializing populations in each sub-system. The Shihmen Reservoir in Taiwan is used as a case study. For comparison, three long-term operation results of a sole GA search and a simulation based on the reservoir rule curves are compared with that of MIGA. The results demonstrate that MIGA is far more efficient than the sole GA and can successfully and efficiently increase the possibility of achieving an optimal solution. The improvement rate of fitness values increases more than 25%, and the computation time dramatically decreases 80% in a 20-year long-term operation case. The MIGA with the flexibility of decomposition strategies proposed in this study can be effectively and suitably used in long-term reservoir operation or systems with similar conditions.     

Differential evolution with subpopulations for high‐dimensional seismic inversion

Seismic inversion has drawn the attention of researchers due to its capability of building an accurate earth model. Such a model will need to be discretised finely, and the dimensions of the inversion problem will be very high. In this paper, we propose an efficient differential evolution algorithm and apply it to high‐dimensional seismic inversion. Our method takes into account the differences among individuals, which are disregarded in conventional differential evolution methods, resulting to a better balance between exploration and exploitation. We divide the entire population into three subpopulations and propose a novel mutation strategy with two phases. Furthermore, we optimise the crossover operator by applying the components having the best objective function values into the crossover operator. We embed this strategy into a cooperative coevolutionary differential evolution and propose a new differential evolution algorithm referred to as a differential evolution with subpopulations. Then, we apply our scheme to both synthetic and field data; the results of high‐dimensional seismic inversion have shown that the proposed differential evolution with subpopulations achieves faster convergence and a higher‐quality solution for seismic inversion.     

An efficiency-improved genetic algorithm and its application on multimodal functions and a 2D common reflection surface stacking problem

Although Genetic Algorithms have found many successful applications in the field of exploration geophysics, the convergence speed remains a big challenge as Genetic Algorithms usually require a huge amount of fitness function evaluations. In this paper, we propose an efficiency-improved Genetic Algorithm, which has both a good global search capability and a good local search capability, and is also capable of robustly handling the premature convergence challenge commonly seen in linear and directed non-linear optimization methods. In our new genetic algorithm, the global search capability is performed via a modified island model, while the local search capability is provided by a novel self-adaptive differential evolution fine tuning scheme. Premature convergence is dealt with via a local exhaustive search method. We first demonstrate the much improved convergence speed of this efficiency-improved Genetic Algorithm over that of our previously proposed advanced Genetic Algorithm on several multimodal functions. We further demonstrate the effectiveness of our efficiency-improved Genetic Algorithm by applying it to a two-dimensional common reflection surface stacking problem, which is a highly nonlinear geophysical optimization problem, to obtain very encouraging results.     

基于自适应遗传算法的MRS-TEM联合反演方法研究

地面磁共振法(MRS)因具有定性、定量分析地下水能力,而备受关注.传统磁共振地层含水量反演多采用均匀半空间模型,忽略电阻率分布信息对结果的影响.针对这一问题,本文基于多层电介质中磁共振响应理论,提出MRS与瞬变电磁(TEM)联合反演方法,通过电阻率分布信息对含水量反演过程的实时修正,提高了解释结果的准确度.反演算法采用自适应遗传算法(AGA)进行,基于繁殖规则,动态调整交叉概率和变异概率,解决了标准遗传算法易未成熟收敛而难以得到全局最优解问题.模型数据表明,含噪10%情况下,联合反演仍能较准确地反映地下含水单元模型结构,对比MRS单独反演优势明显.同时,内蒙古白旗野外观测数据联合反演结果与钻井资料基本一致,充分验证了AGA反演算法的实用性及MRS-TEM联合反演的实际意义.