The paper specifies the fundamentals of stochastic simulation Latin Hypercube Sampling Method (LHS) and presents examples of application of this method for the solution of the stress-strain and stability state of the slope. The stochastic simulation methods enable to involve into the calculation the stochastic character of the input parameters. In the first example the stochastic character of shear strength characteristics of soils is considered. The second example takes into account the stochastic character of the basic parameters of dynamic load (frequency and amplitude of the dynamic load) for the solution of dynamic response (stability factor and displacements) of the slope subjected to surface blasting seismic effect. On the basis of the stochastic analysis we can determine the probability of the model response.
The stress-strain and stability analysis of the slope is the very often solved problem in the geotechnical practice. For the solution of such problem there can be used many computational methods based on limit equilibrium methods (LEQM) or numerical methods (for example finite element method (FEM)), which more or less reflect the various factors determining the behavior of slope. Predicative power of the modeling results is always fundamentally determined by the reliability of the input data. To increase the predicative power of model results (model response) can be used stochastic simulation methods, which are based on the assumption, that the input values of the model are considered as stochastic variables governed by certain laws of probability. Based on the accepted type of the probability distribution of the input parameters of the model and the type of simulation methods the specific values of input parameters are generated. To determine the appropriate random model responses the series of repeated parametric calculations with the generated input values are carried out. Based on the statistical evaluation of the model responses (stochastic variables) the determination of probability range of response results can be obtained.