Inversion Study Of Genetic Algorithm Based On Errors

It's important to determine the fault activity and forecast earthquake that inversing the slip rate of fault. However, the observation data inevitably contain errors and cause results can not be optimal. In view of this question, we think about the data quality control to conduct the inversion and take the Longmen Mountains fault and Yumu Mountains fault as the example. The main works are summarized as follows:1.The conventional algorithms are very much restricted for the non-linear inversion problem. Taking into account the existence of inversion, such as inversion and rely on initial selection easily and other issues fall into the local minimum.This paper focuses on the global optimization algorithm based-genetic algorithms, and applied to fault parameter inversion, which can solve the inverse problem effectively.2.The surface deformation caused by fault movement value is simulated with the dislocation theory. The result of the left-lateral strike-silp-fault is to be observations and be used to computation and analysis with genetic algorithm.The results show that genetic algorithm has good reliability and stability,can be well simulated true value.3.The GPS and other geodetic data inevitably contain errors.In the article we given the mathematical model when the observation data contain errors.4.We process the observation data which contains error, and use it to the inversion taking the Longmen Mountain fault and the Yumu mountain fault as an example. And then we compare the results of the primitive GPS data and the processed GPS data. The result indicated that the vale of the processed GPS data is much stable than the vale of the primitive GPS data. It can reflect well the natureof fault and has good consistency with geological result.