Eldorado - Repositorium der TU Dortmund

Ressourcen aus und für Forschung, Lehre und Studium

Bei diesem Service handelt es sich um das Institutionelle Repositorium der Technischen Universität Dortmund. Hier werden Ressourcen aus und für Lehre, Studium und Forschung gespeichert, erschlossen und der Öffentlichkeit zugänglich gemacht.

Hauptbereiche in Eldorado

Wählen Sie einen Bereich, um dessen Inhalt anzusehen.

Gerade angezeigt 1 - 5 von 9

Fach- und Universitätsübergreifendes 1319
Fakultäten 13860
Graduate School und Graduiertenkollegs 12
Hochschulverwaltung 1249
Sonderforschungsbereiche 1595

Aktuellste Veröffentlichungen

A temperature study of high‐n Rydberg states in Cu2O
(Wiley, 2024-04-21) Heckötter, Julian; Panda, Binodbihari; Brägelmann, Katharina; Aßmann, Marc; Bayer, Manfred
The temperature dependence of Rydberg excitons in Cu_2O with principal quantum numbers n≥ 10 is investigated for bath temperatures between 1.3 and 50 K. The energy shift of Rydberg exciton lines allows us to perform a precise measurement of the Cu_2O band gap as a function of temperature. The Γ^−_5 phonon shows a dominant contribution to the temperature shift of the band gap. The optical properties of Rydberg excitons are analyzed for different temperatures and discussed in the context of phonon scattering as well as thermal ionization of impurities and compared to earlier descriptions in Ref. [1]. The maximum principal quantum number n_max as a function of temperature in crystals of different quality is studied and compared. The observations are correlated to photoluminescence spectra of impurities at different temperatures.
Parametric and Nonparametric Learners for Adaptive Experiments Under Interference
(2026) Hossain, Shaikh Tanvir; Jentsch, Carsten; Jung, Philip
This dissertation develops a framework and a family of both parametric and nonparametric algorithms for the multi-armed bandit (MAB) problem under batched interference. In the first chapter we start with a rigorous overview of the stochastic bandit problem without interference and review the classic Thompson sampling algorithm (Thompson 1933) with the Bayesian regret analysis. In the second chapter, as our first novel contribution, we formalize the stochastic bandit problem under batched interference and define the stochastic bandit, the arm vectors that target different optimals, and the corresponding notions of regret. We then propose four parametric extensions of Thompson sampling, we call them the i.i.d., global-one arm, clustered-one arm, and clustered-multi arm learners, each differing in its strengths across different interference scenarios. Finally, we give some Bayesian regret analysis for the global-one arm learner and end the chapter with numerical simulations that show the strengths and weaknesses of each algorithm. Parametric learners depend on parametric assumptions. In Chapter 3, we propose nonparametric versions of our learners based on Nonparametric Thompson Sampling (Riou and Honda 2020) and the Subsampling Duelling Algorithm (Baudry, Kaufmann and Maillard 2020). These are non-trivial extensions and very flexible algorithms without any parametric assumptions. In the fourth and final chapter we turn to application, we study the problem of learning optimal prices under interference. This is motivated by a real life example of street hawkers in Bangladesh. Here a buyer's purchase decision is influenced by prices offered to other buyers she is socially connected to. We apply both the parametric and nonparametric learners in this setting that maximizes the seller's revenue over a horizon. Together, our contributions open new connections among bandit algorithms, interference, and dynamic pricing that may be useful in practice.
ATMo² Report 1: Mobilitätsverhalten in Nordrhein-Westfalen
(2026-07-01) Kohaupt-Cepera, Kay; Philippi, Lotta; Willen, Sebastian; Opolony, Steffen; Shahbaz Nejad, Bijan; Handte, Marcus
Mobilität in urbanen Räumen steht zunehmend im Spannungsfeld zwischen notwendigen ökologischen Transformationen, begrenzten Verkehrskapazitäten und etablierten Alltagsroutinen. Obwohl in der Rhein-Ruhr-Region vielfältige Alternativen zum privaten Pkw vorhanden sind, bleibt dieser für viele Wege weiterhin das dominante Verkehrsmittel. Das Projekt ATMo² setzt hier an und zielt darauf ab, Reisende durch evidenzbasierte und individuelle Anreize dabei zu unterstützen, vorhandene Alternativen stärker in Betracht zu ziehen und informiertere Entscheidungen bei der Verkehrsmittelwahl zu treffen. Der vorliegende Report beschreibt die Gestaltung der Ausgangsbefragung des Projekts sowie erste Ergebnisse. Die theoretische Grundlage hierfür bilden zwei sich ergänzende Modelle des Mobilitätsverhaltens, die einerseits subjektive Nutzenkalkulationen, Situationswahrnehmungen und Kontextfaktoren sowie andererseits Routinen, Mobilitätssozialisation, Werteorientierungen und soziale Normen berücksichtigen. Auf dieser Basis wurde ein mehrsprachiger Fragebogen entwickelt, der zentrale Dimensionen des Mobilitätsverhaltens, der Verkehrsmittelverfügbarkeit, der Präferenzen und möglicher Verhaltensänderungen erfasst. Die Befragung wurde über verschiedene Online- und Offline-Kanäle verbreitet. Insgesamt konnten 1.174 vollständige Antworten generiert werden; besonders erfolgreich waren dabei die Ausspielung über soziale Medien sowie über das Portal Verkehr.NRW. Die ersten deskriptiven Ergebnisse zeigen eine Stichprobe, die hinsichtlich Alter, Bildungsabschluss und Haushaltseinkommen nach oben verzerrt ist und sich beabsichtigt räumlich deutlich auf Nordrhein-Westfalen konzentriert. Zugleich wird sichtbar, dass den Befragten häufig verschiedene Verkehrsmittel zur Verfügung stehen, der Pkw jedoch über die betrachteten Wegezwecke hinweg weiterhin das meist genutzte Hauptverkehrsmittel bleibt. Damit bestätigen die Ergebnisse die Relevanz eines Ansatzes, der nicht nur die objektive Verfügbarkeit von Verkehrsmitteln, sondern auch Routinen, Wahrnehmungen, Präferenzen und soziale Kontextfaktoren in den Blick nimmt.
Alterations in knee biomechanics and motor performance following 3 months training with the Football+ and 11+ warm-up programs among amateur female players
(Elsevier BV, 2026-03-14) Asgari, Mojtaba; Hägglund, Martin; Terschluse, Benedikt; Sueck, Maximilian; Nolte, Kevin; Schmidt, Marcus; Jaitner, Thomas
Objective This cluster alocated comparative study (DRKS00036644) primarily evaluated the effects of the Football+ and the established FIFA 11+ (the 11+) on knee biomechanical risk factors. A secondary outcome included performance measures and their retention following a 10-week no-intervention period. Methods Three German amateur women's teams (24.3 ± 5.3 years, 1.73 ± 0.07 m, 64.3 ± 8.0 kg) completed the introductory sessions and were team allocated into the Football+ (n = 22), 11+ (n = 19), or control (n = 16) groups. Baseline assessment included a standardized 3D motion analysis of knee biomechanics during single-leg landing and cutting maneuvers and a performance test battery (sprinting, counter movement jump (CMJ), agility, and dribbling speed. Following a 3-month, twice-weekly supervised intervention, post-intervention testing was performed. A 10-week follow-up tested retention of the performance outcomes. Statistical significance was set at α ≤ 0.05. Results ANOVA revealed significant time and time × group interactions across multiple knee biomechanical variables. The Football+ showed consistently larger magnitudes of change compared with the 11+, while the control group demonstrated limited changes (p = .005-0.047; η^2 = 0.14-0.24). Similar interaction effects were observed for sprinting, agility, and dribbling performance (p = .002-0.01; η^2 = 0.13-0.23). Performance improvements were not retained after a 10-week no-intervention period. Conclusion The Football+ and 11+ programs improved high-risk movement patterns associated with an increased risk of knee injury, with more consistent and larger effects observed for the Football+. Improvements in performance measures were observed only following the Football+, supporting its potential as a time-efficient, dual-purpose warm-up. The performance improvements were not maintained following the no-intervention period.
Joining-by-hydroforming of aluminum and poly(ether ether ketone)
(Elsevier BV, 2025-12-30) Weber, Florian; Handge, Ulrich A.; Rakshit, Tanmoy; Dardaei Joghan, Hamed; Hahn, Marlon; Korkolis, Yannis P.; Tekkaya, A. Erman
Joining-by-hydroforming is a process in which components are joined through expansion under internal pressure. Depending on the required fluid pressure and application rate, this process can be technically demanding and challenging to implement on industrial equipment. To address this, a simplified experimental setup was developed to investigate the fundamental joining mechanisms. In this setup, aluminum 6061-T6 (AA6061-T6) and poly(ether ether ketone) (PEEK) rings are force-fitted using a conical punch and segmented conical expansion elements, enabling controlled radial expansion. The resulting assemblies are subsequently separated in a dedicated push-out test. Experimental results show that the required separation force increases with rising elastic strain in the polymer, attributable to an increase in contact pressure according to Coulomb's friction law. This effect diminishes once plastic deformation of the thermoplastic initiates. Furthermore, stress relaxation in PEEK causes a time-dependent decrease in joint strength, reaching a quasi-equilibrium after approximately 10^4 s, as confirmed by relaxation experiments on PEEK coupons. To analyze thermal effects, the entire ring assembly is preheated to defined temperatures in a laboratory furnace. An inverse correlation between joining temperature and joint strength is observed, consistent with the trend identified in the dynamic-mechanical-thermal analysis (DMTA) of PEEK. The proposed experimental method enables rapid identification of the most influential parameters for joining-by-hydroforming, without requiring dedicated hydroforming equipment or production machine time.