Numerical simulations are representations by numerical models. To simulate means to represent a phenomenon by an equivalent system. A simulation transforms the data at the model inputs into one or more results at the model output called inputs-outputs.

Simulation methods use random numbers to vary the random variable (s). This consists of determining the experiments which are in fact an execution of a large number of simulations. We distinguish among these methods those which use a probability law while respecting a limit state of a regional constraint, those which use several probability laws which can be different from one variable to another one by carrying out conditional simulations of Monte-Carlo or those that use the neuronal network.