Reliability Evaluation of Emergent Behaviour in a Flexible Manufacturing Problem.

Scheduling is a key component of manufacturing systems. In general, it describes how multiple jobs or workloads are processed within a limited set of resources to achieve certain manufacturing objectives, such as minimizing the makespan or weighted tardiness. Nevertheless, as the demand for customized and versatile products increased, these systems became more flexible by incorporating flexible manufacturing processes. As a result, it is possible to vary either the machine operations or the product route within highly automated and connected material handling systems. Therefore, the scheduling process is more complex as it involves solving the routing problem within different alternatives and sharing redundant resources between multiple production processes, thereby increasing workload and bottlenecks within the system. Thus, it is imperative to address the scheduling problem while taking into account the physical constraints caused by shared machines and product routing. This paper proposes a Mixed Integer Linear Programming algorithm to solve the flow-shop scheduling problem and minimize the makespan while balancing the machine usage to limit emergent behaviours due to over-utilization. The robustness of the approach is then verified in the case of delays which alter the processing time on machines. Experiments on a limited instance of a flexible manufacturing system confirm the effectiveness of the solution, with a small increase in completion time even for high-magnitude delays.