As economies across the world have become more service orientated, the importance of studying and understanding service operations has steadily increased. Today data analytics and cyber-physical systems are fundamentally changing how services are designed and delivered. Therefore, many companies are exploring new ways to utilize their data to improve the efficiency of their internal processes. This work focuses on service operations in the context of industrial maintenance (i.e. the scheduling of maintenance, repair, and overhaul)
In this domain, the quality of service operations often depends on the ability to manage uncertainties. In practice, companies need to examine, specify, and prioritize all incoming service requests before they can be scheduled—a difficult task which is currently conducted manually by domain experts. In this thesis, offered in cooperation with TRUMPF GmbH + Co. KG, you will explore the feasibility of "mimicking" these human capabilities through machine learning. Based on a unique real-world data set (operational data and machinery data), you will develop a decision support systems for these experts. To do so, you will initially evaluate the general feasibility of using machine learning techniques to predict characteristics of service requests (i.e. required service duration) before they are scheduled. Ideally, you will implement a simple prototype to evaluate the real-world usability and performance of these models with employees.
If you are interested, check the attached PDF document and send your CV, transcript of records, and a brief letter of motivation to firstname.lastname@example.org.