What can you tell us about the topic of your plenary talk in Passau and its importance?

In my presentation, I will be talking about a connection between modeling fairness in optimization problems and bilevel optimization. Fairness is getting increasingly important in our times where we see increasing inequalities in our societies. Thus, policy and decision makers will be more and more interested in integrating fairness into their decision-making process. What we can do then, as operations researchers, is to develop mathematical models that can help them during this process.
In my talk, I will show you a unifiying modeling framework that can address several interesting and practically relevant fairness measures proposed in previous literature.
I will focus on applications in routing, and I will show you how to develop branch-and-cut algorithms to solve this type of problems. The presentation is based on joint work with Alberto Torrejon from the University of Seville, that was part of his PhD thesis, and with his PhD supervisor, Justo Puerto.

What was your own path to operations research, and what attracted you to this field?

I originally studied mathematics and enjoyed virtually every area of the subject. However, it was only quite late in my studies—during my master’s program in mathematics at the University of Belgrade—that I discovered a course in combinatorial optimization. That was the moment I realized I had found a field I was truly passionate about.
After taking that course, I wrote my master’s thesis on Steiner trees and decided to pursue a PhD. Around that time, a PhD position became available at TU Wien. I applied, was fortunate to be selected, and continued working on Steiner trees, as well as other graph optimization problems.
Looking back, that combinatorial optimization course was the turning point that set me on the path to operations research and ultimately shaped my academic career.

You have also worked with industry partners. What can you tell us about those experiences?

These collaborations are both enjoyable and highly rewarding. When working with industry partners, you quickly realize that many of the problems we study in academia are intentionally abstract and often quite far away from the complexity of real-world applications. Practitioners bring a different perspective and introduce many additional considerations that are essential in practice.
As a result, the problems become increasingly complex, often to the point where they appear almost impossible to solve because of the large number of requirements and constraints involved. The challenge is then to identify which aspects are truly essential and which can be simplified, relaxed, or even removed. Finding the right balance between models that are too abstract to be useful and models that are too complex to be tractable is one of the most interesting aspects of applied operations research.
Another challenge is communication. Practitioners do not usually speak the language of optimization or academia, so translating between these different perspectives is an important part of the collaboration. For example, concepts that seem fundamental to us, such as the distinction between objective functions and constraints, are not always framed in the same way by industry partners. Learning to understand each other’s language and priorities is therefore a crucial part of the process. At the same time, this is what makes such collaborations so exciting. There is a great sense of satisfaction in knowing that the methods and results you develop will not simply remain in a research paper or a drawer, but can have a tangible impact and be used to support real-world decision-making.

Do you have any recommendations on what academics could do better to make this interface easier?

I think this is already happening to some extent. In machine learning, for example, interfaces have become increasingly user-friendly, making sophisticated methods accessible to a much broader audience. We are seeing a similar trend in optimization. Companies developing optimization solvers, are continuously working to simplify their interfaces and lower the barriers to adoption.
I believe this is an important step for the growth of our field. Operations research is often perceived as highly technical and difficult to access. By developing tools that are easier to use and by making our methods more approachable, we can help practitioners focus on solving their problems rather than on mastering the underlying technical details. This will make it easier for organizations to adopt optimization methods and benefit from the advances being made in our research community.

You recently wrote a book together with Martin Schmidt and Yasmine Beck. What can you tell us about it?

This project began several years ago when Martin Schmidt and Yasmine Beck started preparing lecture notes for Martin’s course on bilevel optimization at Trier University. More recently, Martin invited me to join the project so that we could expand its scope and incorporate recent developments in mixed-integer bilevel optimization, alongside the topics of linear and continuous bilevel optimization.
Together, we developed the material far beyond the original lecture notes. The book now covers a broad range of topics, including branch-and-cut algorithms, interdiction problems, intersection cuts for mixed-integer bilevel optimization, bilevel optimization under uncertainty, and models with multiple leaders or multiple followers. Our goal was to provide an up-to-date introduction to the field that is useful both for newcomers and for researchers interested in recent advances. A preprint of the book is already freely available on our websites, and the printed version is scheduled to be published by Cambridge University Press at the beginning of next year.

Did you have a specific audience in mind that you would recommend this book to?

The book was written primarily as a textbook. It can be used as the main reference for an introductory course on bilevel optimization, either as an elective course in a master’s program—particularly for students in mathematics, operations research, or computer science—or as an introductory course for PhD students entering the field.
At the same time, the book is intended to be accessible to a broader audience. Researchers from different areas who would like a clear and gentle introduction to bilevel optimization can also benefit from it. We have tried to present both the fundamental concepts and the more recent developments in a way that is approachable for readers who may not already be specialists in the field.
In this sense, I believe the book complements the existing literature quite well. While there are several excellent research monographs on bilevel optimization, this is the first textbook specifically designed to support teaching and self-study at the introductory level.

How do you see the future of OR? What challenges await the OR community?

I think there are really two different questions here: One concerns the challenges within the OR community, and the other concerns how we position ourselves in the broader scientific and industrial landscape.
Within the community, one of the biggest challenges is how we adapt to the rapid progress of AI. We need to understand how these increasingly powerful tools can be integrated into our research and teaching, and how they can help us become more productive. At the same time, we need to think carefully about maintaining scientific quality and originality. Tasks that once required months of work can now sometimes be completed in a fraction of the time with the assistance of AI. This raises important questions about how research will evolve and whether we, as a community, can focus more on asking truly original and fundamental questions rather than simply accelerating established research workflows. The second challenge is how we position operations research externally, particularly in relation to industry. With the growing prominence of AI and machine learning, there is a risk that optimization and OR may be perceived as being absorbed into these broader trends. I believe it is important to communicate clearly that not every decision-making problem can be solved by AI alone. Optimization remains a field with its own expertise, methods, and theoretical foundations.
Of course, optimization plays a crucial role in many AI and machine learning algorithms and applications, but the relationship also works in the other direction. There are many important scientific and practical problems for which optimization methods provide insights, guarantees, and solutions that AI alone cannot offer. Demonstrating this value—to companies, policymakers, and society more broadly—will be essential for the future of our discipline.

When OR 2026 takes place in Passau in September, you will have just started a new position a bit further down the Danube at WU Vienna. What are you most looking forward to about returning to Vienna?

Returning to Vienna really feels like coming home for me. I am very excited to join the Institute of Statistics and Mathematics at WU Wien and to become part of such a vibrant academic environment. What I am particularly looking forward to is building bridges between operations research and the other areas of expertise represented at the institute, especially financial mathematics, statistics, and machine learning. I believe there are many opportunities for fruitful collaborations at the intersection of these fields, and I am eager to explore them.
Beyond that, I am also excited about the broader research environment at WU Vienna. The university is establishing a new department focused on Business Analytics and Decision Sciences, which I expect will create many new opportunities for exchange and collaboration.

Thank you so much for your time and we are looking forward to seeing you in Passau!