The novel reconfigurable modular production system is expected to be the next breakthrough in pharmaceutical manufacturing. In these systems, the target product can be manipulated by changing the modules active on a certain production line, but optimal scheduling is a challenging problem because the on-demand operation of the modular system involves sequence-dependent changeover time/cost and competition of limited modules. Based on the characteristics of the modular production system, this work proposes a multi-stage hybrid discrete/continuous-time scheduling framework to fully utilize the advantages of the two formulations. A discrete-time lot-sizing MILP model is first solved to determine the solution structure, and then an information mapping algorithm extracts the necessary information from the discrete-time solution. Finally, a continuous-time LP model is constructed and solved based on the extracted information. The proposed model shows decent results in its application to a reconfigurable modular continuous production system with seven API products.

新型的可重构模块化生产系统有望成为制药行业的下一个突破。在这些系统中，可以通过更改某个生产线上活动的模块来操纵目标产品，但是最佳调度是一个具有挑战性的问题，因为模块化系统的按需操作涉及与序列有关的转换时间/成本以及有限竞争的竞争。模块。基于模块化生产系统的特征，这项工作提出了一个多阶段混合离散/连续时间调度框架，以充分利用这两种公式的优点。首先求解离散时间批量MILP模型，以确定解决方案的结构，然后信息映射算法从离散时间解决方案中提取必要的信息。最后，基于提取的信息构造并求解连续时间LP模型。所提出的模型在将其应用于具有七个API产品的可重构模块化连续生产系统中时显示出了可观的结果。