AI-based Decision-making Model for the Development of a Manufacturing Company in the context of Industry 4.0

Managers are looking for solutions that will be helpful when deciding on the purchase of new technologies, in order to adapt the enterprise to the Industry 4.0 concept. Nowadays, many approaches suitable for smart manufacturing systems involving maintenance workers are based on Artificial Neural Networks (ANN). This paper presents an approach to measuring the effectiveness of the use of an IT system supporting the realisation of business processes in the maintenance department and describes the empirical research results of maintenance workers (121) within Polish manufacturing companies with automotive branches. Finally, this paper seeks to integrate the first two main research results and ANN, into a novel, decision-making model regarding the implementation of activities and investments aimed at increasing the level of a company's automation. The architecture of ANN classifier was chosen in a series of experiments. The Levenberg-Marquardt method and genetic algorithms were used in training process. The performance of the classifier was measured using the sum of squared errors and the error function with the regularisation term in the form of the sum of squared norms of Jacobian matrices. The best performing classifier achieved 95.8% accuracy on the test dataset.