Leverage synergy by cyber-physical systems for the convergence of the eco systems mobility, infrastructure and energy in the circular economy for the Society 5.0

Project background

The background to the Cynergy4MIE project lies in the increasing demands for sustainable and efficient production processes in Europe. Industrial companies are under pressure to optimize their production methods in terms of resource conservation, energy efficiency and environmental compatibility, while at the same time increasing productivity.

Project objective

As part of the project, THRO will develop innovative production processes in which people and machines work together sustainably and efficiently. The aim is to use reinforcement learning to create optimal production plans that both improve resource utilization and increase energy efficiency. To this end, a virtual training simulation is being set up that makes it possible to train reinforcement learning agents to optimize sustainable production processes.

In addition, a holistic training framework is being implemented that maps all relevant elements of production - from intralogistics and material and machine dynamics to different product requirements. Seamless integration into existing ERP and MES systems is ensured in order to make the developed methods usable in real production. Finally, the simulated conditions are compared with real production environments in order to identify optimization potential and continuously develop the system further.

Project procedure

The methodology of the project is based on the combination of modern technologies such as reinforcement learning and simulation with traditional industrial systems in order to develop sustainable and efficient production processes.

An essential component is the development of a virtual training environment in which RL agents are trained. This simulation models real production scenarios and enables the behaviour and dynamics of these elements to be tested and optimized in a controlled environment.

After training, the developed system is integrated into real production environments. Finally, the methodology is completed through extensive testing and validation in real production environments. The comparison of the simulated conditions with real production processes serves to close the so-called "reality gap", i.e. to identify the differences between simulation and reality and to further improve the system.

Innovation

The use of reinforcement learning in production planning enables dynamic optimization of the use of resources, which leads to more efficient and sustainable processes. The close integration of productivity targets with sustainability requirements such as the reduction of energy consumption and emissions is particularly innovative. The seamless integration of these new technologies into existing systems such as ERP and MES platforms ensures that companies can integrate the innovations directly into their production processes. This practice-oriented approach helps companies to reduce costs, increase their competitiveness and achieve their sustainability goals.