events and seminars

MSc Seminar on February 23, 12.45-13.30

The speakers will be:


Anatoliy Babic: “Game State Evaluation in Football”


Maarten Otten: “Stratified Breast Cancer Follow-Up Using a Continuous-state POMDP”

For a list of MSc seminars, click here .

Healthcare Operations Research Summer School in Twente
July 21 – 25, 2017
University of Twente, Enschede, the Netherlands

We are excited to announce a special summer school on Healthcare Operations Research this summer at the University of Twente, the Netherlands! The event targets PhD students working on OR applications focused on improving healthcare. Participants will learn state-of-the-art OR techniques taught by renowned international professors, learn about how to successfully engage with the healthcare practice, and meet other PhD students from around the world. The program includes a wide variety of components including technical lectures, tutorials, inspiring seminars, interactive activities and a design competition of a serious game that demonstrates a healthcare OR/OM concept to healthcare professionals. The total workload of the summer school is 2 European Credits (2 ECs = 56 hours), consisting of preparation and contact hours. For more information and registration please follow this link.


For a list of seminars and conferences, click here.

Dinsdag Seminar

Dinsdag Seminar on December 13, 12.45-13.30, HB 2A

Error bounds for stationary performance of random walks in the quarter plane based on inhomogeneous perturbations


Xinwei Bai




Random walks in the quarter plane with homogeneous transition probabilities are considered in this work. Given a non-negative reward function on the state space, we are interested in the expected stationary performance. Due to the difficulty of direct derivation of this performance for general random walks, upper and lower bounds are constructed based on the performance of a perturbed random walk for which the stationary probability distribution is a sum of geometric terms.

We consider inhomogeneous transition probabilities for the perturbed random walks, which means that the transition probabilities are different at every state. The bounds are constructed using the Markov reward approach. In the end, an explicit expression for the error bound is given. The error bound result does not depend on the values of the inhomogeneous transition probabilities. Therefore, only the existence of those probabilities is needed.

Numerical experiments indicate that inhomogeneous perturbed random walks can give better error bounds than perturbations based on homogeneous random walks. The reason is that by allowing for inhomogeneous transition probabilities, we don’t require the pairwise structure between geometric terms that is necessary for homogeneous random walks.

Dinsdag Seminar on Thursday November 19, 12.45-13.30 in Carre 3D

For a list of Dinsdag seminars, click here.