Parameters Estimation Of Vti Media, Using Hybrid Genetic Algorithms

At this work is proposed a sequential method based on hybrid genetic algorithms (HGA) composed by the genetic algorithm and the Nelder mead simplex algorithm to estimate parameters of a model from seismic reflexion. The model represents layers into underground. In this way, the parameters of a layer form a chromosome of the HGA and the set of layers forms a seismic model. In the inversion process of multiple layers, the representative velocities are the main problem to get the layer velocities. Consequently, the proposed method estimates sequentially the seismic parameters for solving this problem. The experiments use seismic models with horizontal layers with isotropy and vertical transverse isotropy (VTI). The proposed method is compared to the method of Medeiros (2005); that considers a chromosome for all seismic models. The results, when is applied the proposed method, are better than refereed method. The data with the travel times are synthetic and without noise. Other advantage is the reduction of the runtime of estimation algorithms; that allows processing of more quantity of data.