KI-basierte Methoden zur Optimierung von Produktionsprozessen

Abstract

Motivation and problem definition The ongoing digital transformation makes artificial intelligence (AI) a central component of modern production systems. AI-based methods offers potential to increase efficiency, quality and costeffectiveness in industrial processes, but their implementation often fails due to technical, organizational and personnel hurdles such as a lack of data standards, heterogeneous IT systems or a lack of competencies (C. Lee et al., 2025; Wuest et al., 2016). ). International comparisons show, that companies in the DACH region lag behind countries such as the USA or China in the use of AI in production (BMWI, 2020; Statista, 2025). There is a need for a systematic analysis of AI methods and a practice-oriented framework to support companies in the introduction and further development of data-driven processes. This thesis addresses this gap by investigating AI technologies, areas of application, prerequisites and risk-benefit aspects in industrial production. Content structure and methodology The thesis is divided into nine chapters, which are structured along a three-pillar structure: (1) Basics of industrial process optimization (e.g., Lean, Six Sigma), (2) theoretical introduction to AI methods (e.g., machine learning, deep learning), and (3) Use of AI in production. A systematic literature review (Chapter 5) forms the methodological basis, in which scientific databases with clearly defined search terms and inclusion/exclusion criteria were analysed (Page et al., 2021). The results were systematically evaluated in Chapter 6 and transferred to a maturity model with five levels (Initial to Transformative) and five dimensions (Technology, Organization, Data, Competencies, Strategy) in Chapter 7. Chapter 8 discusses the findings in the context of the research questions and derives practical recommendations for action. The work focuses on industrial processes, while operational business processes such as controlling or human resources are excluded. Concrete results of the work The analysis identifies machine learning, deep learning, digital twins, and hybrid approaches as core AI technologies for production optimization, supported by robust IT/OT integration (Mo et al., 2023; Schmitz et al., 2025). Areas of application such as quality assurance, predictive maintenance and production control are established, with untapped potential in adaptive systems, especially for SMEs (Babic et al., 2021; Raffin et al., 2022). Success factors include standardized data platforms, clear governance, and interdisciplinary competencies (Kober et al., 2024). A risk-benefit assessment shows, that AI increases efficiency and quality at higher maturity levels, while risks such as high costs or interpretation problems can be minimized by XAI and security measures (Thawon et al., 2025; Zhao et al., 2025). The developed maturity model provides guidance for companies to assess and further develop their AI capabilities, highlighting IT/OT integration as the key to data-driven optimization. Future research should explore empirical validations and SME-specific adoption pathways.

Date of Award	2025
Original language	German (Austria)
Supervisor	Sonja Straßer (Supervisor)