自适应免疫克隆粒子群算法的地震波阻抗反演
|
摘要
针对粒子群优化算法应用于地震波阻抗反演问题时易陷入局部极小值和计算量大的问题,提出一种自适应免疫克隆粒子群(AICPSO)算法。该算法引入免疫机制,根据个体浓度和适应值概率定义了个体置换算子,能够避免粒子群算法陷于局部极值。为避免由进化过程中大量相同抗体引起的算法退化现象,根据记忆库和抗体群不同的特性,采用自适应变异算子更新记忆库,而依据抗体浓度和亲和度更新下一代抗体群体。经数值模拟和实际波阻抗资料反演表明,该算法不依赖于初始模型,收敛速度快且结果可靠。免疫粒子群优化算法为解决地震波阻抗反演问题提供了一条可行途径。
An adaptive immune clonal particle swarm optimization(AICPSO) algorithm was presented to overcome the shortcoming of converging to local optimum and of massive computation the PSO algorithm is apt to fall when applied in seismic wave impedance inversion.The local extremum problem of PSO could be avoided in this algorithm due to defining the individual substitution operator,which is determined by fitness of individual and density probability based on imported immune mechanism.To prevent the algorithm degradation caused by a large number of the same antibodies during evolution,the memory base was modified by adaptive mutation operator and the swarm of antibodies of next generation was modified according to antibody concentration and affinity degree based on different properties of them.The proposed AICPSO is numerically simulated as well as being applied into actual seismic wave impedance inversion,and the data results show that it is not influenced by initial model,converges fast and has reliable result.AICPSO provides an effective means for seismic wave impedance inversion.
An adaptive immune clonal particle swarm optimization(AICPSO) algorithm was presented to overcome the shortcoming of converging to local optimum and of massive computation the PSO algorithm is apt to fall when applied in seismic wave impedance inversion.The local extremum problem of PSO could be avoided in this algorithm due to defining the individual substitution operator,which is determined by fitness of individual and density probability based on imported immune mechanism.To prevent the algorithm degradation caused by a large number of the same antibodies during evolution,the memory base was modified by adaptive mutation operator and the swarm of antibodies of next generation was modified according to antibody concentration and affinity degree based on different properties of them.The proposed AICPSO is numerically simulated as well as being applied into actual seismic wave impedance inversion,and the data results show that it is not influenced by initial model,converges fast and has reliable result.AICPSO provides an effective means for seismic wave impedance inversion.
引文
[1]易远元.地震波阻抗反演的粒子群算法实现[J].石油天然气学报,2007,29(3):79-81.
[2]戚玉涛,刘芳,焦李成.基于信息素模因的免疫克隆选择函数优化[J].计算机研究与发展,2007,45(6):991-997.
[3]李阳阳,焦李成.量子免疫克隆多目标优化算法[J].电子与信息学报,2008,30(6):1368-1371.
[4]戚玉涛,刘芳,焦李成.求解大规模TSP问题的自适应免疫归约算法[J].软件学报,2008,19(6):1265-1273.
[5]公茂果,焦李成,杜海峰.用于约束优化的人工免疫响应进化策略[J].计算机学报,2007,30(1):37-46.
[6]高鹰,谢胜利.免疫粒子群优化算法[J].计算机工程与应用,2004,23(6):30-33.
[7]廖建坤,叶东毅.基于免疫粒子群优化的最小属性约简算法[J].计算机应用,2007,27(3):550-552.
[8]陈颖,徐晓晖.基于免疫克隆原理的改进粒子群优化算法的研究[J].系统仿真学报,2008,20(6):1471-1474.
[9]胡江强,郭晨.启发式自适应免疫克隆算法[J].哈尔滨工程大学学报,2007,28(1):234-239.
[10]朱奇光.利用免疫克隆粒子群混合算法实现自适应PMD补偿[J].计算机工程与应用,2007,43(26):345-349.
[11]徐果明.反演理论及其应用[M].北京:地震出版社,2003:112-116.
[12]姚姚.地球物理反演———基本理论与应用方法[M].武汉:中国地质大学出版社,2002:72-82.
[13]张文博.新型自适应免疫克隆混合算法及应用研究[D].吉林:吉林大学出版社,2007.
[14]LI Y,Wu T.A Novel Immune Algorithm for Complex OptimizationProblems[C]//Proceedings of the 5th World Congress on IntelligentControl and Automation,2004:2279-2283.
[15]尚荣华,焦李成.免疫克隆算法求解动态多目标优化问题[J].软件学报,2007,18(11):2700-2711.
[16]Ho S L,Yang S,Ni G,et al.Aparticle swarm optimization-based meth-od for multi-objective design optimizations[J].IEEE Trans Magn,2005,41(5):1756-1759.
[2]戚玉涛,刘芳,焦李成.基于信息素模因的免疫克隆选择函数优化[J].计算机研究与发展,2007,45(6):991-997.
[3]李阳阳,焦李成.量子免疫克隆多目标优化算法[J].电子与信息学报,2008,30(6):1368-1371.
[4]戚玉涛,刘芳,焦李成.求解大规模TSP问题的自适应免疫归约算法[J].软件学报,2008,19(6):1265-1273.
[5]公茂果,焦李成,杜海峰.用于约束优化的人工免疫响应进化策略[J].计算机学报,2007,30(1):37-46.
[6]高鹰,谢胜利.免疫粒子群优化算法[J].计算机工程与应用,2004,23(6):30-33.
[7]廖建坤,叶东毅.基于免疫粒子群优化的最小属性约简算法[J].计算机应用,2007,27(3):550-552.
[8]陈颖,徐晓晖.基于免疫克隆原理的改进粒子群优化算法的研究[J].系统仿真学报,2008,20(6):1471-1474.
[9]胡江强,郭晨.启发式自适应免疫克隆算法[J].哈尔滨工程大学学报,2007,28(1):234-239.
[10]朱奇光.利用免疫克隆粒子群混合算法实现自适应PMD补偿[J].计算机工程与应用,2007,43(26):345-349.
[11]徐果明.反演理论及其应用[M].北京:地震出版社,2003:112-116.
[12]姚姚.地球物理反演———基本理论与应用方法[M].武汉:中国地质大学出版社,2002:72-82.
[13]张文博.新型自适应免疫克隆混合算法及应用研究[D].吉林:吉林大学出版社,2007.
[14]LI Y,Wu T.A Novel Immune Algorithm for Complex OptimizationProblems[C]//Proceedings of the 5th World Congress on IntelligentControl and Automation,2004:2279-2283.
[15]尚荣华,焦李成.免疫克隆算法求解动态多目标优化问题[J].软件学报,2007,18(11):2700-2711.
[16]Ho S L,Yang S,Ni G,et al.Aparticle swarm optimization-based meth-od for multi-objective design optimizations[J].IEEE Trans Magn,2005,41(5):1756-1759.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心