PhD Position Statistical Integer Linear Programming in High Dimensions (TU Delft)

PhD Position Statistical Integer Linear Programming in High Dimensions (TU Delft)

This PhD position lies in the intersection of statistics (high-dimensional techniques) and optimization (integer linear programming) to handle the problems raised by big data. We are looking for a PhD candidate who can develop theoretical and practical tools for analysing and solving complex large-scale optimization problems with data-driven constraints.

Many real-life optimization problems can be modelled and solved using integer linear programming (ILP). However, the input parameters needed for such an ILP are subject to uncertainty because of, for example, estimation errors or unexpected disturbance. In this project, the case is considered where the uncertain input data is high dimensional, which means that the number of available samples of our data is of the same order or even less than the dimension of the uncertain input parameter. For that reason, new statistical techniques need to be introduced for efficiently handling estimation errors in ILPs with high-dimensional input parameters that combine the theory of random matrices, modern shrinkage methods and machine learning.
The developed tools will be applied on problems arising in transportation. When solving flow-based routing problems, part of the input data is a time-varying origin-destination (OD) matrix, which gives the number of trips between each origin and destination. These matrices often include many OD pairs, but for each time step, only a limited number of observations is available. Therefore, the challenge in these routing problems is to design a robust network given the uncertain high dimensional OD-matrices.

MSc degree in Mathematics, Computer Science or Statistics/Econometrics.
Good writing and presentation skills in English, and certainly proficient in programming.
Initiative, drive and ability to setup, organise and execute your research.
An affinity with teaching and guiding students.
Good organisational skills and ability to work independently and in a team.
Strong interest in high dimensional statistics and integer linear programming.
The PhD candidate will be jointly supervised by Dr. ir. Theresia van Essen (Optimization) and Dr. Nestor Parolya (Statistics).

For information about this vacancy, you can contact Theresia van Essen, assistant professor (Optimization), email:, or Dr. Nestor Parolya, assistant professor (Statistics), email:
For information about the selection procedure, please contact Miriam Heemskerk, MSc., HR-Advisor, email:
Application procedure

To apply, please mention vacancy number TUD00339 and send your motivation letter (approx 1 page), CV (incl. publication list if available) and list of academic transcripts (BSc and MSc) before September 1, 2020. You can send your application material to: Make sure to include in your motivational letter (and/or CV) examples of projects in which you successfully demonstrate your skills relevant to this project. We would appreciate if you could also mention when you could start.

More about TU Deflt, our department and this position can be found here

Zurück zur Übersicht