• search hit 1 of 20
Back to Result List

Optimizing production processes via improved resilience and state-of-the-art AI technologies

  • Disruptions regularly hit economies. Scholars and industry experts suggest many strategies to avoid disruptions or handle them effectively. Two things are repeatedly mentioned: Increasing resilience and deploying artificial intelligence (AI) technologies. In this dissertation, we1 look at both aspects and focus our efforts on production processes. To get a well-rounded view, we apply various research methods, i.e., surveys, case studies, and systematic literature reviews (SLRs). In our first paper (Chapter 2), we investigate the perceived organizational resilience of companies in the German manufacturing industry. We perform an SLR to analyze existing research on organizational resilience measures. We see that existing (qualitative) resilience measures are complex, challenging to interpret, and therefore, hard to scale and apply across multiple industries. Based on this, we develop a novel, low-threshold resilience measure consisting of six resilience items about the past perceived internal/external resilience, current perceived internal/external resilience, and anticipated need for internal/external resilience, called the Enterprise Resilience Index (ERI). Finally, we conduct an empirical study with ~200 German experts across various industries. Our survey shows that the German manufacturing industry perceives itself as relatively resilient, with significant differences between industries and company sizes. We also see that they anticipate a high need for external resilience across industries in the future. Most strikingly, the Machinery industry shows the lowest ERI levels while it anticipates a relatively high need for resilience, showing the development need for this industry in terms of resilience. To explore the aspect of AI, we focus on waste incineration plants (WIP) in Chapter 3. WIPs have various levels of automation, but they still rely on manual operations by human operators. Consequently, the combustion process is managed rather inefficiently, and steam outputs and emission levels are not optimal. Thus, we investigate how reinforcement learning (RL) can help enhance process automation and thus optimize the combustion process, e.g., by making more frequent and diverse interventions. An RL agent is trained via trial and error with a reward function that includes the optimization criteria. Since the actual equipment, i.e., the real WIP, cannot be used as the training environment, a digital twin is built using original plant data and a neural network. The RL agent is then trained in this offline environment with the deep Q-network algorithm (DQN). Our work demonstrates that a digital twin of a WIP can be built in a data-driven way. We show that the RL agent outperforms the human operator, increasing the steam output by 7.4% and reducing the oxygen level by 3.6%. Thus, applying RL might benefit the plant operator financially due to increased output and the environment in terms of reduced emission levels. Finally, we look at a practical aspect of AI: AI readiness and adoption (Chapter 4). Many companies across various sectors have adopted AI technologies. However, the supposedly high adoption rates are misleading since many applications are rather experimental and not applied in key business areas. We believe that this limited AI adoption arises from a lack of AI readiness. We conduct a case study in the waste incineration industry with over 160 clients and investigate which strategies facilitate AI adoption in not-AI-ready industries. Based on these interactions, we distill five strategies that counter typical AI readiness barriers, thus increasing AI readiness: education, trust, customer centricity, focus, and collaboration. These strategies focus on transforming businesses just as much as necessary to prepare them for the AI technology that is supposed to be implemented. With increased AI readiness, chances for AI adoption rise. We are convinced that these strategies can be applied in various environments. In summary, this dissertation gives empirical evidence and expands the literature on organizational resilience and benchmarking, reinforcement learning and digital twins, and AI readiness and adoption. 1 The term “we” refers to the authors of the respective chapters, as noted at the beginning of each chapter.

Download full text files

Export metadata

Additional Services

Search Google Scholar Statistics
Metadaten
Author:Martin Schlappa
URN:urn:nbn:de:hbz:992-opus4-9774
Place of publication:Vallendar
Referee:Stefan Spinler, Liji Shen
Document Type:Doctoral Thesis
Language:English
Date of Publication (online):2024/04/26
Date of first Publication:2024/04/26
Publishing Institution:WHU - Otto Beisheim School of Management
Granting Institution:WHU - Otto Beisheim School of Management
Date of final exam:2023/04/20
Release Date:2024/04/26
Tag:Deutschland; Fallstudie; Künstliche Intelligenz; Organisationelle Resilienz; Produktionsprozess
Artificial intelligence; Case study; Germany; Organizational resilience; Production process
Page Number:78
Institutes:WHU Supply Chain Management Group / Chair of Logistics Management
Licence (German):Copyright this PhD thesis