256869 research outputs found
Sort by
Optimierung des Campus-IT-Service Management: Entwicklung eines Jira-Projektkatalogs auf der Grundlage der Bedürfnisse der Universität und der Folgenabschätzung
No full textEffzientes IT Service Management (ITSM) im Hochschulkontext verlangt einen Ausgleich zwischen der starren Ausrichtung an Frameworks wie Information Technology InfrastructureLibrary (ITIL) und der dezentralen, dynamischen Struktur des akademischen Betriebs. An der TU Wien nutzt die Campus IT Jira als zentrales Werkzeug für das Servicemanagement.Das Fehlen einer einheitlichen Implementierungsstrategie hatte jedoch zur Folge, dass Konfigurationen ad hoc vorgenommen wurden, was zu uneinheitlicher Serviceerbringung und schwankender Zufriedenheit der Nutzenden führte. Diese Arbeit begegnet diesen Problemen durch die Konzeption eines modularen Jira-Projektkatalogs,der darauf abzielt, Service-Desk-Prozesse zu standardisieren und zugleich den heterogenen Anforderungen von Studierenden, Lehrenden und Verwaltungseinheiten gerecht zu werden.Methodisch verfolgt die Arbeit einen Mixed-Methods-Ansatz, um den Ist-Zustand desITSM an der TU Wien zu erfassen und die vorgeschlagene Lösung zu bewerten. Zunächst liefert eine quantitative Auswertung historischer Jira-Daten eine Leistungsgrundlage.Diese zeigt, dass die durchschnittliche Bearbeitungszeit für Standardanfragen mit 0,9Tagen zwar effzient ist, eine Reihe komplexer Fälle jedoch die Gesamtservicequalität überproportional beeinträchtigt. Ergänzend dazu liefert eine Stakeholder-Befragung qualitative Einsichten in die Anforderungen der Nutzenden und macht das Spannungsfeld zwischen dem Wunsch nach klar strukturierten Prozessen und dem Bedürfnis nach operativer Freiheit deutlich.Aufbauend auf diesen Ergebnissen entwickelt die Arbeit einen modularen Projektkatalog,der standardisierte Projektvorlagen und Workflows bereitstellt, ohne dabei ein starres Einheitsmodell vorzugeben. Dieses Konzept versetzt die Campus IT in die Lage, Verwaltungsvorgaben einzuhalten und dennoch die von unterschiedlichen Organisationseinheiten der Universität benötigte Flexibilität zu gewährleisten. Darüber hinaus präsentiert die Arbeit einen Proof of Concept (PoC) für ein Wissensabrufsystem, das Retrieval-Augmented Generation (RAG) einsetzt, um die Arbeitslast der Support-Mitarbeitenden weiter zu verringern und die Effzienz bei der Bearbeitung von Vorfällen zu steigern.Die Evaluierung belegt, dass der vorgeschlagene Katalog den identifizierten betrieblichen Engstellen gezielt entgegenwirkt. Durch die Ausrichtung der technischen Konfiguration an den empirisch ermittelten Anforderungen der Stakeholder, schafft der Katalog eine skalierbare Grundlage für die Weiterentwicklung der IT-Servicebereitstellung.Effcient ITSM in higher education requires balancing the rigid governance of frameworks like ITIL with the decentralized, flexible nature of academic environments. At TU Wien, the central Campus IT department relies on Jira for service management, but the absence of a standardized implementation strategy has led to ad hoc configurations,inconsistent service delivery, and varying levels of user satisfaction. This thesis addresses these challenges by developing a tailored Jira Project Catalogue designed to standardize service desk operations while accommodating the diverse needs of students, faculty, and administrative staff.The research employs a mixed methods approach to evaluate the current state of ITSMat TU Wien and validate the proposed solution. First, a quantitative analysis of historical Jira data establishes a performance baseline, revealing that while the median resolution time for standard requests is effcient (0.9 days), a “long tail” of complex issues disproportionately impacts overall service quality. Second, a stakeholder survey captures qualitative insights into user requirements, highlighting the tension between the need for structured processes and the desire for operational autonomy.Based on these findings, this thesis proposes a modular Jira Project Catalogue that provides standardized project templates and workflows without enforcing a universal constraint. This design allows Campus IT to maintain governance standards while offering the flexibility required by different university departments. Additionally, there search introduces a PoC for a knowledge retrieval system utilizing RAG, aimed at further reducing agent workload and improving incident resolution.The evaluation demonstrates that the proposed catalogue directly addresses the identified operational bottlenecks. By aligning technical configuration with empirical stakeholder needs, the catalogue offers a scalable foundation for optimizing IT service delivery.The results suggest that a modular, needs based approach to ITSM configuration can significantly enhance both operational effciency and user satisfaction in a complex university environment
Simulation-based generation of heuristics for decision-making in stochastic environments
No full textDecision-making in stochastic environments often requires a trade-off between performance and interpretability. Although Reinforcement Learning (RL) excels at creating adaptive policies, the resulting solutions are not transparent. Conversely, while heuristics offer transparency, they often lack optimality and adaptability. In this work, we present a general framework that combines the strengths of both approaches. First, we use RL to train a policy on a Markov Decision Process (MDP). Then, we extract transparent heuristics in the form of decision trees via interpretable learning (VIPER). To conclude our method, we apply pruning to the tree, aiming to simplify its structure and improve the generalisation of the resulting rule set. We demonstrate this approach using a logistics case study involving significant variability in production and demand. The resulting heuristics outperform expert-designed rules and match the performance of the original RL policy, offering transparency and robustness. This method allows for data-driven, explainable decision-making that does not require domain-specific expertise
EDT-SaFL: Semi-Asynchronous Federated Learning for Edge Digital Twin in Industrial Internet-of-Things
No full textThrough conducting equivalent model training within the paradigm of edge intelligence, the Digital Twin Edge Networks (DITEN) have been widely employed in the Industrial Internet-of-Things (IIoT) to facilitate the cost-effective execution without the operational disruption. However, due to the insufficient consideration of heterogeneity in computing and communication capabilities of distinct industrial terminals in the Digital Twin (DT) model training, the existing approaches of DT construction/update have unbalanced model training cost and loss in the whole life cycle of DT model, hindering the abilities of quick responding to complex and dynamic productions and ensuring the data consistency of virtual-real space. To address this issue, we define a global loss minimization problem with constraint, and propose an original approach of semi-asynchronous federated learning, named EDT-SaFL, as a promising solution. Considering the collaborative utilization of heterogeneous resources, and the contribution of local data quantity and quality to the global model update, the EDT-SaFL consists of three important operations, Terminal Selection for Model Training, Self-Adaptation of Local Training Iterations, and Semi-asynchronous Global Aggregation. With the analysis of convergence, complexity and communication overhead, the experiments have evidently demonstrated the superiority of EDT-SaFL on the datasets of CIFAR-10 and Industrial-Equipment
Thermal and athermal nucleation of MgSi co-clusters in Al-Mg-Si alloys
Get PDFThe interplay between thermal and athermal nucleation of MgSi co-clusters during quenching of solution-heat-treated Al-Mg-Si alloys is investigated through computer simulations. Thermal nucleation is typically described by classical nucleation theory, which refers to the formation of supercritical nuclei via the diffusion-controlled attachment of solute atoms to clusters of critical size. In the process of athermal nucleation, pre-existing subcritical nuclei become supercritical due to a decrease in critical size, for instance, as a result of increased undercooling during quenching. In this study, we develop a comprehensive nucleation model that integrates thermal and athermal contributions, offering new insights into the MgSi co-cluster formation in Al-Mg-Si alloys during continuous cooling. The results reveal that athermal nucleation is the predominant nucleation mechanism for MgSi co-clusters during quenching. Furthermore, the dependencies of thermal and athermal nucleation on cooling rate, temperature, and alloy composition are elucidated
Stochastic very weak solutions to parabolic equations with singular coefficients
No full textA class of stochastic parabolic equations with singular potentials is analyzed within the chaos expansion framework, utilizing the Wick product to handle the multiplication of generalized stochastic processes. The analysis combines the chaos expansion method from white noise analysis with the concept of very weak solutions from partial differential equation theory. The stochastic very weak solution to the parabolic evolution problem is defined, and its existence and uniqueness are established. For sufficiently regular potentials and data, we demonstrate the consistency of the stochastic very weak solution with a stochastic weak solution. An illustrative example is provided, potential applications are reviewed, and future challenges are outlined
Classifying different criteria for learning algebraic structures
No full textIn the last years there has been a growing interest in the study of learning problems associated with algebraic structures. The framework we use models the scenario in which a learner is given larger and larger fragments of a structure from a given target family and is required to output an hypothesis about the structure's isomorphism type. So far researchers focused on Ex-learning, in which the learner is asked to eventually stabilize to the correct hypothesis, and on restrictions where the learner is allowed to change the hypothesis a fixed number of times. Yet, other learning paradigms coming from classical algorithmic learning theory remained unexplored. We study the ‘‘learning power’’ of such criteria, comparing them via descriptive-set-theoretic tools thanks to the novel notion of E-learnability. The main outcome of this paper is that such criteria admit natural syntactic characterizations in terms of infinitary formulas analogous to the one given for Ex-learning in. Such characterizations give a powerful method to understand whether a family of structures is learnable with respect to the desired criterion
Corrigendum to “Multi-neighborhood simulated annealing for the oven scheduling problem” [Comput. Oper. Res. 177 (2025) 106999]
No full textThe Oven Scheduling Problem (OSP) is an NP-hard real-world parallel batch scheduling problem that arises in the semiconductor manufacturing sector. It aims to group compatible jobs in batches and to find an optimal schedule in order to reduce oven runtime, setup costs, and job tardiness. This work proposes a Simulated Annealing (SA) algorithm for the OSP, encompassing a unique combination of four neighborhoods and a construction heuristic as initial solution. An extensive experimental evaluation is performed, benchmarking the proposed SA algorithm against state-of-the-art methods. The results show that this approach consistently finds new upper bounds for large instances, while for smaller instances, it achieves solutions of comparable quality to state-of-the-art methods. These results are delivered in significantly less time than the literature approaches require. Additionally, the SA is extended to tackle a related batch scheduling problem from the literature. Even in this case, the algorithm confirms its effectiveness and robustness across different problem formulations by improving results for many instances.
Graphical abstrac
Computational Mechanics Approaches for Stiffness and Strength Estimates of Plant-Based Biocomposites
No full textThis article summarizes and reviews recent advancements in computational mechanics applied to wood and plant-based biocomposites. Emphasizing sustainability, these materials are gaining importance due to their excellent mechanical properties and low environmental impact. The paper highlights the complex hierarchical microstructure of wood and biocomposites, spanning from critical molecules like cellulose to macroscopic features, all of which influence mechanical performance and are considered in various modeling approaches. A range of computational modeling techniques is reviewed, including homogeneous, mesoscale, and microscale models, as well as molecular dynamics simulations. Homogeneous models directly address macroscopic failure mechanisms, while mesoscale models are able to capture the effects of material structures and provide insights into the interactions of a material’s constituents. Microscale approaches, for example, investigate the structural behavior of cellulosic fibrils, whereas molecular models examine atomic-level interactions. Despite significant progress in recent years, challenges remain in bridging scales, accounting for environmental factors such as moisture, and optimizing computational efficiency. This article offers a comprehensive overview of computational methods across these scales and provides an understanding of the strengths and limitations of each approach
Impact of doping and channel inhomogeneities on the stability of industrially fabricated WS₂ FETs
No full textWe report doping-dependent charge trapping in WS₂ field-effect transistors fabricated on a 300 mm wafer. In particular, higher n-type doping–associated with smaller channel areas–correlates with an increased density of active defects. This behavior explains the asymmetric threshold voltage degradation observed in large-area ambipolar devices, where the n-branch consistently shifts more than the p-branch under gate bias stress (by a factor of ~ 3). Through electrical characterization and photoluminescence mapping, we attribute this asymmetry to process-induced inhomogeneities in the WS₂ layer and its chemical environment, which lead to enhanced n-type doping at the channel center relative to the edges. The non-uniform doping profile and conduction of the 2D channel are then captured using an equivalent circuit model that quantitatively reproduces the observed degradation asymmetry and corroborates our interpretation. These results have important implications for the development of large-scale 2D semiconductor transistors, highlighting the impact of unintentional process-induced doping and channel heterogeneity on device performance and reliability