Immune Genetic Seismic Residual Static Correction Based on High Order Cumulant

摘要

Estimation of residual static correction in noise-contained data is a nonlinear, inverse problem of time delay. The method described in this paper uses the fourth-order cumulant of the data and models a parametric estimation for the time delay, combining genetic algorithm (GA) with artificial immune (AI) to perform global optimization of the model. The proposed immune genetic algorithm (IGA) focused on two aspects: maintaining the population diversity by using the information entropy theory; controlling generality by antibody Concentration. Numerical simulations demonstrate that the method can accurately estimate time delay, even when the non-Gaussian signal buried in Gaussian noises.