The inverse analysis in geotechnical engineering is a typical complicated nonlinear function optimization problem. To solve this problem

the global optimization algorithm is a very good method. In order to offset the disadvantages of low searching efficiency using traditional methods of back analysis

the particle swarm optimization and genetic algorithm are hybridized to construct a global optimization algorithm according to their compatibility and algorithm complementary. A new evaluation function method is established which unites multiobjective function to single objective function according to the actually measures and corresponding calculated results. The mixed penalty function method is adopted to transform constraint condition to unconfined condition and a new objective function is presented. Based on the improved particle swarm optimization

the program of back analysis is coded

in which ABAQUS is embedded as a module in the optimization algorithm. Through a tunnel model as an example

the proposed inverse program is verified; and the results show that this method is a very good inverse analysis method and its efficiency is very good. So the proposed method is a feasible method for back analysis of stress

seepage and displacement in geotechnical engineering.