Probabilistic optimization via approximate p-efficient points and bundle methods

摘要: For problems when decisions are taken prior to observing the realization of underlying random events, probabilistic constraints an important modeling tool if reliability is a concern. A key concept numerically dealing with that p-efficient points. By adopting dual point view, we develop solution framework includes and extends various existing formulations. The unifying approach built on basis recent generation bundle methods called on-demand accuracy, characterized by its versatility flexibility. Numerical results for several difficult confirm interest approach.