Information technologies Optimization of the Connection Weights and Thresholds in the Seismic Inversion Neural Network Algorithm

In the seismic inversion model, as the neural network algorithm there are some problems, the convergence bad, accuracy is not high. This paper presents a seismic inversion based on artificial fish Swarm Optimization neural network models. First to initialize fish mapping of chaos optimization and ergodicity of artificial fish-swarm search and adaptive strategies of artificial fish-swarm algorithm of optimization search strategy, and using as a parameter change measure, in the course of operation of the algorithm Adaptive adjustment of parameters, and finally building seismic inversion based on artificial fish Swarm Optimization neural network model. Simulation experiments show that improved artificial fish-swarm algorithm presented in this paper has better accuracy optimization, artificial fish-swarm algorithm of neural network based on improved convergence properties better and seismic inversion algorithm based on improved neural network model of meanvariance changes more slowly.