Abstract: Production planning and scheduling are two of the most important decision-making problems in supply chain optimization. To ensure the efficiency of decision-making, planning and scheduling were integrated, with consideration of demand uncertainty. For a multi-period production planning and scheduling problem, planning and scheduling deterministic models were established in each period firstly, and they were integrated through production correlation. Then, demand uncertainty was introduced and decision variables were represented with a finite number of scenarios in a two-stage stochastic programming model. At last, rolling horizon strategy was used to achieve consistency between planning and scheduling results in an iterative process. The case study demonstrated that compared with the conventional method, the stochastic programming method could reduce total cost under demand uncertainty. Combined with the hierarchical integration method, production operational and economic optimization was achieved.

Key words: planning and scheduling, uncertainty, stochastic programming, optimization, model, systems engineering

摘要： 生产计划与调度是化工供应链优化中两个重要的决策问题。为了提高生产决策的效率，不仅要对计划与调度进行集成，而且要考虑不确定性的影响。对于多周期生产计划与调度问题，首先在每个生产周期内，分别建立计划与调度的确定性模型，通过产量关联对二者进行集成。然后考虑需求不确定性，使用有限数量的场景表达决策变量，建立二阶段随机规划模型。最后运用滚动时域求解策略，使计划与调度结果在迭代过程中达到一致。实例结果表明，在考虑需求不确定性时，与传统方法相比，随机规划方法可以降低总费用，结合计划与调度的分层集成策略，实现了生产操作性和经济性的综合优化。

关键词: 计划与调度, 不确定性, 随机规划, 优化, 模型, 系统工程