Cognition-enhanced, Self-optimizing Production Networks


Design a viable and changeable production management system to improve the dynamical adaption to optimal operating points
phase 2

(1) Verify and validate the cybernetic reference model of self-optimizing production management with consideration of human decision making and integrating the perspectives of production and quality management
(2) Build test-beds for experimental research in a real production environment and develop prototypes of cybernetic solution components
(3) Establish a comprehensive demonstrator to enable future integrative research beyond the traditional boundaries of disciplines


The Research Area focuses on the management systems and principles of a production system. It aims at controlling the complex interplay of heterogeneous processes in a highly dynamic environment. In the first funding period of the CoE, a reference model for production management, based on the Viable System Model of Stafford Beer, that serves for a starting point for investigations in the sub-project, was developed. The FIR is working together with the Institute of Industrial Engineering and Ergonomics (IAW), the Human Computer-Interaction-Center (HCIC) and the Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen University and contemplates the three subject areas “Corporate Information- and Material-Flow”, “Self-optimizing Production Planning and Control (PPC)” and “Cognitive Production Line”.

  Copyright: © Marc_Winkel_2015

Corporate Information- and Material-Flow

In the context of the simulations model advances could be achieved. In particular the model was extended by additional influencing parameters such as dynamic minimum inventory level. Moreover the supply chain was examined regarding interdependencies of different variables. To develop and validate the results a second logistic demonstrator was developed during the project year 2015. The logistics demonstrator which was developed in collaboration with partners of the RWTH Aachen Campus Cluster Logistics and the FIR illustrates possibilities and potentials of horizontal and vertical integration.

In different baseline experiments, which aimed to investigate socio-technical factors, the influence of user-diversity, information complexity and amount, and the user interface on decision efficacy and efficiency in material disposition decisions was identified and quantified. Furthermore a second version of the “Q-I-game” called “LogisticSim“ was developed.


Self-optimizing Production Planning and Control (PPC)

The aim of the PPC case is to develop a self-optimizing PPC to meet the requirements of frequent checks and adjustments at low manual effort. In 2015, the second prototype called WoPS+ was transferred into WoPS 4.0 by changing the database structure to allow faster and further analyses. The possibilities to include machine downtimes and detailed shift schedules were integrated. Furthermore, algorithms for pattern recognition of actually applied sequencing rules were implemented. To allow further and easier configuration of simulation scenarios, the uploaded production program can now be varied on the user interface. The already existing benchmark database was ameliorated by programming a customized benchmark configuration onto the user interface.


Cognitive Production Chain

In order to study human-robot interaction, various simulation studies were conducted and the results were validated in the cognitive assembly cell at the WZL. The results showed that the prediction time and the mental effort could be decreased by using human-like movements. The validation of the graph-based assembly sequence planner was completed successfully by further simulation studies that targeted the reduction of the number of human-robot changes for the purpose of risk reduction.

Within the scope of developing a reference architecture for cloud-based condition, monitoring the pull-off process of a packaging machine was used for an example implementation. Therefore a Matlab/Simulink-model was built analytically. The validation of the model was conducted using the experimental process data of a tubular bag machine at the WZL. Additionally, various machine learning methods were tested and compared with respect to the identification of the condition of the components.


Based on the current results, an interactive demonstrator for the ICD D1 is currently being developed in collaboration with the ICD "Virtual Production Intelligence". This displays self-optimizing systems between all company levels (shop floor to supply chain). The overall D1 demonstrator allows users to experience the relationships between all levels of the organization itself and to realize the benefits of self-optimization. Further operations in the various cases refer to the extension of WoPS 4.0 platform through the addition of functionality to improve the input data using algorithms, further studies of the effect of support systems and business intelligence on the efficiency and effectiveness of decisions, the development of the game “LogisticSim” as well as the construction of a workplace for direct human-robot collaboration. Moreover joint publications in Cyber Physical Systems started with the institute-cluster IMA, ZLW and IFU.

Find out more:

Technical Demonstrator "Corporate Material and Information Flow"

Tecnical Demonstrator "Cognitive Assembly Cell"

Technical Demonstrator "Self-optimizing Production Planning and Control"