Treating the cavern group as a special system

the corresponding relationship between the system and the mechanical parameters of the rock mass is obtained by using the parallel evolutionary neural network FEM. The worst combination of parameters related to the stability of surrounding rock mass is searched out by genetic algorithms (GA). Through sensibility analysis

the synthetical evaluation of the effects of mechanical parameters of soft rock mass on the stability of cavern group is given. Moreover

the key rock mass could be determined to provide the design and construction with further instructive suggestions. The application shows that the method is reasonable and it has the advantages of intelligent and synthetical analysis.