Quoika, Vivian

Technology regulation and data quality considerations demand a higher control over company data. In this literature review, we synthesize a data governance taxonomy which emphasize the evolving challenges associated with managing diverse data modalities, including numerical tabular data, big data, images, videos and textual content for learning algorithms. This systematic literature review collects foundational concepts, theoretical frameworks and organizational structures, highlighting the critical roles of stakeholders of governance principles and of policy developments to synthesis a comprehensive taxonomy including the data modalities. The analysis underscores the importance of a tailored governance approach that address modality-specific issues such as metadata management, privacy and security. Technological and methodological considerations, including data quality management, lifecycle policies as well as interoperability and standardization. Our study combines knowledge management and considerations about data modalities which are especially relevant for general artificial intelligence and provides a robust foundation for advancing both theoretical understanding and practical implementation of effective data governance. The paper contributes to a robust data governance and aims at advancing theoretical understanding as well as practical implications for quality management across heterogeneous data environments and which creates insight for policy makers.

This computational study investigates a production scheduling problem for a two-stage hybrid flow shop (HFS) with parallel machines at least at one stage and with zero inter-stage waiting policies between process steps. This scenario is important in industries such as steel production and chemical processing, where cooling time between process steps must be avoided. In this paper, we propose eight dispatching rules that are applied to instances of up to 200 jobs, and benchmark them based on various performance metrics that demonstrate the effectiveness of the proposed heuristic approaches. The dispatching rules, such as Shortest Task Time (STT) and Shortest Processing Time (SPT) are combined with machine assignment rules, such as First Available Machine (FAM) and Minimum Idle Time (MIT), and optimized for makespan and total completion time in the no-wait HFS. Furthermore, for our computational study, we investigate two sequencing approaches - stage-oriented decomposition (A1) and reduction to a flow shop problem (A2) - in the benchmark of this computational study.