New VDMA guideline Self-learning production processes: Capturing Machine Learning Step by Step
The new VDMA guideline supports the topic of machine learning and shows step by step how the introduction can take place in your own company.
Machine Learning is an important technology for the implementation of Industry 4.0. As part of Artificial Intelligence, Machine Learning brings exciting and new approaches to the optimization of products and processes for machine and plant construction.
Reinforcement method for machine learning An
interesting machine learning method is reinforcement learning, which is based on the principle of reward and means learning by trial and error. It is particularly promising for the control of industrial processes, and is particularly suitable for learning intelligent control strategies in an industry 4.0 context.
Real time as central argument
With the help of reinforcement learning, control strategies can be learned both for very complex processes and for complex environmental conditions of machines, without having to model them explicitly. Another advantage of reinforcement learning is the ability to determine a control strategy in real time, while simulation would be too computationally intensive.
"Our application example at AZO showed that machine learning processes have the potential to control complex processes more efficiently", says Dieter Herzig, CDO of AZO GmbH Co. KG and chairman of the VDMA working group accompanying the project.
New guide supports the introduction of Machine Learning
In cooperation with the Institute for Enterprise Cybernetics e. V. (IfU) from Aachen and a VDMA Industrial Working Group accompanying the project, the VDMA Industry Forum 4.0 has developed the guideline "Self-learning production processes - Introduction strategy for reinforcement learning in industrial practice".
This was presented to the public for the first time at an event today and was concretized with experiences from mechanical engineering.
The aim of the guideline is to tackle the introduction of machine learning on a broad scale step by step and thus provide VDMA members with practical guidance. A tool is provided for the development of an own implementation strategy and basics and terminology are explained. Predefined guiding questions and toolboxes help.
About the Industry 4.0 Forum
The VDMA has bundled its internal association know-how in the Industry 4.0 Forum. The forum consists of an interdisciplinary team of VDMA experts. As partners, they offer support to the member companies as well as the professional associations and departments of the VDMA in the relevant fields of action.