Mixed-model sequencing versus car sequencing: comparison of feasible solution spaces
比较混合模型排序（MMS）和汽车排序（CS）两种方法在汽车装配线排序中的可行解空间，发现MMS能生成更多可行序列，但实际中CS可能因特定规则产生更多不同序列。
In the automotive industry, a great challenge of production scheduling is to sequence cars on assembly lines. Among a wide variety of scheduling approaches, academics and manufacturers pay close attention to two specific models: Mixed-Model Sequencing (MMS) and Car Sequencing (CS). Whereas MMS explicitly considers the assembly line balance, CS operates with sequencing rules to find the best car sequence fulfilling the assembly plant requirements, like minimising work overload for assembly workers. Meanwhile, automakers including Renault Group are increasingly willing to consider other requirements, like end-to-end supply chain matters, in production planning and scheduling. In this context, this study compares MMS- and CS-feasible solution spaces to determine which workload-oriented sequencing model would be the most appropriate to later integrate new optimisation. We introduce two exact methods based on Dynamic Programming to assess the gap between both models. Numerical experiments are carried out on real-life manufacturing features from a Renault Group assembly plant. They show that MMS generates more feasible sequences than CS regardless of the sequencing rule calculation method. Only the sequencing rules used by real-life production schedulers result in a higher number of distinct feasible sequences for CS, highlighting that the plant might select a sequence with work overload situations.