The improvements of natural cognation exclusion strategy for genetic algorithm and its application in residual static correction
Pan Shu-lin1, Wu Bo2, Gao Lei3
1. Faculty of Resources and Environment,Southwest University of Petroleum,Chengdu City,Sichuan Province,610500,China;
2. Deyan Branch,Research Institute of Exploration,Southwest Branch Company of Sinopec,Deyang City,Sichuan Province,618000,China;
3. Faculty of Computer Science&Technology,Southwest University of Petroleum,Chengdu City,Sichuan Province,610500,China
Abstract:The premature convergence often appears when using existing genetic algorithm in residual static correction,resulting in difficult realization of the global optimization,the key reason is that the selection method of model male parent is too simple,leading to the population diversity of the model male parent which was used in the cross-operation is too low.Regarding the problem,based on the model male parent group which was generated by the simulated annealing probability function the selection operator which uses natural cognation exclusion strategy replaces the matching priority selection operator in existing genetic algorithm.The theoretical model and field data experiments show that the method is superior than the conventional genetic method,its main representation is that the natural cognation exclusion strategy has the characteristics of the assurance of group population diversity,the inhibition of premature convergence,the expansion of the search space,and the realization of the global optimization.
潘树林, 吴波, 高磊. 自然血亲排斥策略对遗传算法的改进及其在剩余静校正中的应用[J]. 石油地球物理勘探, 2010, 45(6): 802-806.
Pan Shu-lin, Wu Bo, Gao Lei. The improvements of natural cognation exclusion strategy for genetic algorithm and its application in residual static correction. OGP, 2010, 45(6): 802-806.