Stochastic nonlinear complementarity problems: stochastic programming reformulation and penalty-based approximation method (Q965063)

The authors consider a class of stochastic nonlinear complementarity problems in a finite-dimensional space. By using a suitable residual, they convert it into a constrained stochastic optimization problem and after into a simple constrained problem via a penalty method. Some conditions which provide equivalence for these problems and for their stationary points are given. A sample based approximation of the penalized optimization problem is proposed and some numerical experiments are described.