Relaxations for probabilistically constrained programs with discrete random variables (Q1197883): Difference between revisions

The author suggests a new method providing an outer convex approximation of the (generally nonconvex) set of feasible solutions typical for stochastic linear programs with joint probability constraints and random discretely distributed right-hand sides. The relaxed problems are of the form of linear programs and their optimal values serve as bounds for the optimal value of the original stochastic program.