Special Issue: AI and Data Driven Decisions in Manufacturing

This is a special issue on the topics of the conference MIM 2022 (https://hub.imt-atlantique.fr/mim2022/). Following their presentation at the conference, some authors were invited to submit an extended version of their work to this special issue. However, the special issue is also open to papers that were not presented at the conference if they are in the scope of the issue.

Despite the explosion of sensors and data generated from the shop floor, advanced planning and process control systems suffer from the absence of well-integrated solutions that deliver the full potential of digitalization. How to make efficient decisions with this plethora of data? In practice, very often, decision-makers and operators rely mainly on their experience. The challenge now is to integrate the latest advances in analytics and machine learning into manufacturing decision-support technologies to fully benefit from AI and digitalization on the manufacturing shop floor.

Manufacturing companies are increasingly adopting information technologies to have a complete and accurate view of their processes on the shop floor. Such technologies are often referred to as the "internet of things"(IOT), and they include: RFID, various automatic sensors on machines and products, manufacturing execution systems, among others. Technologies to collect and store the IOT data are available, and this data is extremely valuable since it can yield better decisions at each level (strategic, tactical, operation, control). The data available nowadays in manufacturing systems allow the integration of AI technics in all manufacturing decisions: process design, planning, scheduling, and production execution control. From one hand, this data can be used to improve the accuracy of decision models for process planning, production planning, and scheduling. This includes the automatic learning of constraint programming and mathematical programing algorithms, as well as the inclusion of uncertainty within these models. On the other hand, machine learning algorithms may be trained based on this data to improve the quality of the processes, and to control the production.

In this context, this special issue welcomes original theoretical approaches and new applications of AI in manufacturing and data driven approaches for manufacturing decisions:

• Data driven approaches for process planning, production planning and scheduling.

• Knowledge graph for the integration of manufacturing data from different sources.

• Adaptable optimization models for manufacturing decisions.

• Digital twins and data driven simulation for planning and scheduling.

• Optimization approaches for shop floor scheduling and control.

• Integration of machine learning and mathematical programming.

• Optimization under uncertainty: stochastic and robust optimization.

• Machine learning based algorithms for quality control.

• Human detection approaches for human robot collaboration.

• Explainable artificial intelligence.

• Human-centric decision support systems.

All papers should adhere to the Journal of Intelligent Manufacturing aims and scope and publication criteria. Methods on the interface of optimization, simulation and artificial intelligence as well as papers dealing with real-life industrial context and technologies are particularly welcome.