Dr. Mirjam Dür of the Johann Bernoulli Institute for Mathematics and
Computer Science at the University of Groningen has been awarded a
prestigious NWO Vici grant of € 1.5 M. The Vici grants are the most
substantial grants in the NWO Innovational Research Incentives Scheme.
The subsidy is directed at promising young researchers who have
completed their doctorate less than 15 years ago and who have shown that
they have the ability to successfully develop their own innovative lines
of research.
Dr. Dür is a Rosalind Franklin Fellow and an Assistant Professor of
Mathematical Optimization at the University of Groningen. Her research
deals with developing algorithmic methods to compute optimal solutions
of mathematical optimization problems.
Optimization problems pop up in many practical applications, for example
railway scheduling, telecommunication networks, personnel staffing,
finance, but also in fields like the engineering sciences or molecular
biology.
A typical problem in discrete optimization requires choosing the best
solution from a large but finite set of possibilities, for example, the
best order in which a salesperson should visit clients so that the route
travelled is as short as possible. The difficulty is that the number of
possibilities is usually very large, so that enumerating them one by one
is not an option. A similar challenge occurs with problems that exhibit
a large number of local optima while the practical application requires
knowledge of the global optimum. This typically happens with models
involving quadratic or other nonlinear functions.
In the last decades, mathematical methodologies to treat discrete models
have seen an enormous progress, and so have techniques to deal with
nonlinear models. However, as more and more practical applications
become tractable, more and more problems appear which involve both
discrete and continuous variables. For those problems, the necessary
theory and algorithms are not yet satisfactory.
The Vici grant has been awarded for research on problems that involve
both these features, discrete variables and nonlinear functions. The
results of the project will lead to new powerful solution methods for
classes of optimization problems that have up to now not been solvable
in reasonable time.
See Mirjam Dür's personal website http://www.math.rug.nl/~mirjam/
for more information.
