Identification of geo-material rheological constitutive model based on fast-convergent genetic algorithm

Abstract

To identify rheological constitutive model of geo-materials, one generalized constitutive law is applied. So, the problem of model identification is transformed to the problem of traditional parameters identification. According to the relationship of objective function and optimization methods, the global optimization method, such as evolutionary algorithm, is very suitable to solve parameter identification problems. A new fast-convergent genetic algorithm is applied in this study. In this new algorithm, there are only two individuals in one population. So, the whole computation efficiency of optimization back analysis will be very high. Using this new back analysis method, a real engineering example of one underground coal mine roadway is used to verify the computing ability of the algorithm to real problems. The results show that the efficiency of optimization back analysis can be improved greatly with this new algorithm.