Academic personal website of
Dr. Sven Mallach

Welcome!

This website gives a brief overview on my research, (third-party funded) projects, and professional background with further references. For current teaching, please visit the website of my working group. If you like to contact me, feel free to do so via [firstname].[lastname]@uni-siegen.de.

Academic Background & Appointments

As of September 2025, I am an interim professor for Network and Data Science Management at the University of Siegen, Germany. With a neatless transition, I led the chair of Management Science at the University of Siegen before and since April 2024.

Since 2020, I am also a research fellow at the University of Bonn, Germany, where I have been affiliated with the High Performance Computing and Analytics Lab until my engagement in Siegen. Moreover, I am elected to the steering committee of the transdisciplinary research area Mathematics, Modelling and Simulation of Complex Systems at the University of Bonn.

Prior to these appointments, I have been an interim full professor for Discrete Optimization as well as a postdoc at the Department of Mathematics & Computer Science at the University of Cologne. In 2015, I received a PhD (Dr. rer. nat.) from the Department, working with Michael Jünger. My diploma is in Computer Science, received from the Technical University of Dortmund, Germany, in 2008.

PhD Students

Besides my small team in Siegen, and thanks to my granted DFG project, Chris Cohadari is my PhD student at the University of Bonn since 2025.

Research

I am working in mathematical programming and operations research, on the edges between and with a variety of applications in quantitative economics, computer science, and mathematics. When asked to cut down the ultimate goal of my research into a short sentence, then it is to push frontiers in the solution of practically relevant optimization problems. This means in particular to increase the size or share of problem instances that can be solved routinely, by means of better algorithms and models.

To achieve this goal, I frequently develop and combine methods and ingredients that lead to significant computational advances and that have their foundations in:

• Discrete Optimization, in particular (Mixed-)Integer Linear Programming and Binary Quadratic Programming
• Continuous Optimization, especially Linear, (Convex) Quadratic, and Semidefinite Programming
• Combinatorial Optimization, e.g. Combinatorial Algorithms, Polyhedral Theory, and Graph Theory
• Data and Network Analysis, Data Science, and Machine Learning
• Algorithm Engineering, Efficient Data Structures, and High Performance Computing

More concretely (but non-exhaustively), such methods are e.g. sophisticated branch-&-cut (or outer approximation) algorithms, reformulation and linearization techniques, and engineered separation algorithms. I publish on both general (broadly applicable) methods as well as to tailored (problem-specific) ones, see also projects and publications.

I contribute to and have running projects on applications in a broad mix of disciplines, for instance in logistics and transportation, production and manufacturing, health-care, and the automotive industry, and I'm always interested in and quickly adapt to new applications. On a meta-level, I gained a particular expertise in ordering, assignment, scheduling and binary quadratic problems which are ubiquitous.

Selected Lines of Research

This a selection from my major research directions and projects with recent developments and publications.

1. Sophisticated Models and Algorithms for Applications

This line of research focusses on the solution of various combinatorial optimization problems that arise as economic and interdisciplinary applications. Specifically, it is my passion to design and improve mathematical models with emphasis on their better practical solution, and to develop sophisticated solution techniques that go far beyond “plugging a model into a solver”. Methods of mine have been applied in business practice, e.g. by Siemens for manufacturing projects. Currently, I run projects on further industrial as well as health-care applications. The most recent publications deal with:

• The Target Visitation Problem which models route planning applications with additional preferences between the locations to visit, beyond the classical travel distance or cost. The paper has been presented at Symposium on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS) 2025.
• Asymmetric Betweenness Problems with applications in facility layout and relations to the more general Quadratic Linear Ordering Problem (this is the most recent preprint).

2. Quadratic Unconstrained Binary Optimization (QUBO) and the Maximum Cut Problem (MaxCut)

I work on exact methods for these two strongly related problems from various perspectives. Thereby, an already tremendous and steadily increasing amount of applications from various disciplines is formulated and solved as a QUBO.

• Integration and common enhancement of solution techniques from Polyhedral Combinatorics, Integer Linear Programming, and Semidefinite Programming in a collaboration with Angelika Wiegele. A corresponding DFG project has been granted and started in 2025.
• For MaxCut, I combine classic theory and novel insights to obtain Integer Programming Formulations based on Spanning Trees. The corresponding paper has been presented at the ISCO 2024.
• Techniques for the separation of facet-defining odd-cycle inequalities, giving a leap to exact solvers: One and Two
• Exact solvers for sparse instances based on Integer Programming, like McSparse
• The Spin Glass Server
• A new instance library with meta data information, see a prototype here

3. General Methods for Constrained Binary Quadratic Optimization

These works are generally applicable to a wide range of applications, e.g.\ involving assignment or matching constraints.

• I develop the theory and work on the computational effectiveness of the Inductive Linearization Technique for binary programs with linear constraints and a quadratic objective function, quadratic constraints, or both. Effectively, it generates a (often small) subset of the constraints defining a first-level RLT of the original program which - by construction - suffice to linearize it without any further effort.
• In 2023/2024, I worked out an inexact simplex algorithm that performs quadratic optimization over the vertices of (0-1) polyhedra.

Publications

A full publication list (i.e., with theses, recent articles to appear, as well as preprints) is here.
A list of peer-reviewed publications is available via ORCID.
Finally, you might also take a look at DBLP, however, as you see some of the rather math-oriented journal publications are either not listed there or only listed with a quite large delay.