Sample-path optimization of convex stochastic performance functions

摘要: In this paper we propose a method for optimizing convex performance functions in stochastic systems. These can include expected static systems and steady-state discrete-event dynamic systems; they may be nonsmooth. The is closely related to retrospective simulation optimization; it appears overcome some limitations of approximation, which often applied such problems. We explain the give computational results two classes problems: tandem production lines with up 50 machines, PERT (Program Evaluation Review Technique) problems 70 nodes 110 arcs.