Maria Vlasiou is a Full Professor at the University of Twente, The Netherlands, an Associate Professor at the Eindhoven University of Technology (TU/e), and a Research Fellow of the European research institute EURANDOM. 

Her research centres on the performance of stochastic processing networks with layered architectures, focusing on energy networks, high-tech manufacturing, and stochastic performance analysis. Other interests include Lévy processes, perturbation analysis for heavy-tailed risk models, large deviations for non-monotone stochastic recursions, and proportional fairness in heavy traffic for bandwidth-sharing networks. She has supervised nine PhD theses on these topics.

Her research has been funded by grants exceeding 4.5M from more than 10 science foundations, universities, societies, and organisations. She is the co-author of more than 45 refereed papers, the co-recipient of the best paper award in ICORES 2013, the Marcel Neuts student paper award in MAM8, and of the 3^rd prize of the 8th conference in Actuarial Science.

Prof. Vlasiou is associate editor in four journals and has refereed for about 50 international journals, conferences, and national science foundations. She serves as the representative of the Netherlands at the IMU Committee for Women in Mathematics and at the European Women in Mathematics association. She is additionally in the advisory boards of the national associations of women in mathematics in Cyprus and in Greece.

Stochastic processes driven by non-convex resource allocation problems

The project focuses on stochastic processes of which the dynamics are driven by resource allocation problems related to non-convex optimisation problems. These processes naturally occur as multi-scale models that describe congestion in man-made systems such as energy and communication networks. Being Markov processes on infinite-dimensional state spaces, their analysis is mathematically challenging and current techniques insufficient: Linear approximations of the driving resource allocation problem can lead to a poor understanding of such systems and to suboptimal designs, especially under critical loading. We develop mathematical results in the context of electric vehicle charging, analysing their fluid models and applying them to derive control rules that handle congestion in electricity networks efficiently.

Extreme-Value Theory for Large Fork-Join Queues

The project is inspired by modeling delays in supply chains for high-tech manufacturers, such as ASML, Philips Healthcare, and Boeing. These supply chains are very large compared to the typical supply chain in industry. A common property is that many high-tech suppliers specialize in producing and delivering a specific component of the final product. In this system, the slowest supplier determines the delay of the manufacturer.

To model this delay, we consider the N-server fork-join queueing network, in which each server represents a unique supplier, and the arrival stream denotes orders from the manufacturer. We investigate the behavior of the longest queue and the longest waiting time by proving limiting results as the number of servers N converges to infinity. We additionally propose centralized base-stock and capacity policies to minimize costs incurred by delays. To achieve these objectives, we use results from extreme-value theory, diffusion approximations, large deviations principles, theory on heavy-tailed random variables, and newsvendor problems.

Maria Vlasiou teaches courses in Probability, Stochastic processes, infectious-disease epidemiology, and performance modelling. She regularly supervises BSc, MSc, and PhD theses, internships, and PDEng projects. She is also a Bachelor College coach and a student mentor. 

Prof. Vlasiou has obtained her UTQ diploma in 2010 and has continued her training in education through various professional development courses. She has served in the departmental Didactics committee and UTQ committee and has obtained several grants on research in mathematics education.