概要

The paper aims to construct an efficient optimization method for identifying creep and destructuration related governing parameters of soft structured clay. An elastic viscoplastic model has been developed and adopted. Different optimization processes, by genetic algorithm or particle swarm optimization with uniform or random samplings initialization methods, are carried out to obtain the material parameters from conventional undrained triaxial tests performed on a K ₀-concolidated natural soft clay. All comparisons demonstrate that the uniqueness of the solution is better guaranteed with the genetic algorithm rather than with the particle swarm optimization method. Furthermore, the efficiency of genetic algorithm has been verified by simulating other tests.