Stability analysis of one stage stochastic mathematical programs with complementarity constraints (Q415375)

One stage stochastic programming problems with complementarity constraints are considered. The aim is to study the quantitative stability of the solution sets, optimal value, and M-stationary points with respect to the varying probability measure. The following properties of the mentioned characteristics as functions of the probability measure are proved: the optimal solution set is upper semi-continuous, the optimal value is Lipshitz continuous, and M-stationary points are upper semi-continuous. For the stochastic programming problem of the type of sample average approximation it is proved that the optimal value and M-stationary points converge exponentially to their true counterparts with probability one when the sample size increases. The latter result can be useful for the development of numerical algorithms different from the standard Monte Carlo schemes developed earlier for the considered problems.