An Efficient Trajectory Method for Probabilistic Production-Inventory-Distribution Problems

摘要: We consider a supply chain operating in an uncertain environment: The customers' demand is characterized by discrete probability distribution. A probabilistic programming approach adopted for constructing inventory-production-distribution plan over multiperiod planning horizon. does not allow the backlogging of unsatisfied demand, and minimizes costs while enabling it to reach prescribed nonstockout service level. It strategic that hedges against undesirable outcomes, can be adjusted account possible favorable realizations quantities. modular, integrated, computationally tractable method proposed solution associated stochastic mixed-integer optimization problems containing joint constraints with dependent right-hand side variables. concept p-efficiency used construct finite number trajectories, which turn are employed solve constraints. complement this idea designing preordered set-based preprocessing algorithm selects subset promising p-efficient trajectories. Finally, resulting disjunctive problem, we implement special column-generation limits risk congestion resources chain. methodology validated on industrial problem faced large chemical turns out very efficient: finds minimal integrality gap provides substantial cost savings.