Market Demand Oriented Data-Driven Modeling for Dynamic Manufacturing System Control

In this paper, a market demand driven modeling framework is developed. The market demand is modeled as an end-of-line virtual machine based on which the market demand dissatisfaction (MDD) can be measured as production loss using event-based analysis. A general Markovian continuous-flow model is developed for market demand-driven systems with multistage production networks combining manual and automatic processes. Machine failure bottlenecks (MF-BNs) and machine capacity bottlenecks (MC-BNs) are defined and identified based on event-based indicators. A supervisory control algorithm is integrated in the framework to reduce MDD and improve system productivity through identification and mitigation of MF-BNs and MC-BNs. Simulation-based analysis will also be utilized, and case studies are performed to validate the effectiveness of the modeling framework and the supervisory control policies.