Analysis accuracy and robustness of parameters inversion in probability integral method by genetic algorithm

This paper focuses on accuracy and robustness of parameters inversion in probability integral method by genetic algorithm. For this, uniform design experimental method, subsidence prediction software and genetic algorithm program are used. Result shows that parameters in probability integral method can be retrieved precisely by genetic algorithm with relative errors of the retrieved parameters are less than 1.5%; retrieving parameters by genetic algorithm has a great applicability in different areas such as measuring errors, gross errors, loss of observation stations, etc.