Hard combinatorial problems arise in many areas of computing science
and its applications. I am interested in algorithmic methods for solving
these problems as efficiently and robustly as possible.
In particular, I study a fundamental approach called
stochastic local search (SLS), which underlies many of
the best-performing algorithms for combinatorial
Much of my work on SLS algorithms is focused
on the propositional satisfiability problem (SAT)
and its optimisation variant, MAX-SAT,
two well-known and conceptually simple combinatorial
Besides developing better SLS algorithms for these and other problems,
my research is focussed on
obtaining an improved understanding of SLS behaviour and
performance, which is often achieved by means of advanced
empirical analysis techniques.
These techniques form the basis for
the area of empirical algorithmics,
which complements and extends established theoretical methods
for the analysis of algorithms with principles and techniques
based on an empirical science approach.
Most of my bioinformatics research is concerned
with computational methods for biomolecular
structure prediction and design,
with a focus on RNA and DNA secondary structure.
Together with my students and collaborators,
I have developed high-performance algorithms
for the following problems:
I am also actively interested in the role of RNA structure
in splicing, in regulatory interactions and networks,
gene expression analysis and motif discovery.
- RNA secondary structure design,
- RNA secondary structure prediction with pseudoknots, and
- DNA word design and HP protein structure prediction.
Computational Intelligence (aka AI):
Most of my work in computational intelligence is closely
related to my research in empirical algorithmics,
and focused on solving hard AI problems, including
propositional satisfiability (SAT) and maximum satisfiability
(MAX-SAT), constraint satisfaction problems (CSPs),
as well as scheduling problems.
I have also done some work on linear planning,
and I have been involved in work preference representation
and reasoning using ceteris paribus networks (CP nets).
Furthermore, I have recently started working in an area
I call human-centred information management (HCIM),
which is concerned with methods and systems for
organising and retrieving information, such as text documents
or audio content, in a way that is intuitive and efficient
for human users.
Research in this highly interdisciplinary area combines
different aspects of computing science and music; my work
is mainly on the following topics:
I am one of the originators of
GUIDO Music Notation,
a rich, human-readable symbolic music representation
I was also the director of the
an academic effort for studying algorithmic aspects in music
launched in 1993 at Darmstadt University of Technology.
In the context of this project, our team developed the
SALIERI System, a powerful interactive software environment
for the score-level manipulation of musical material.
- composition, variation and analysis of music;
- music representation issues and formalisms;
- music programming languages and environments;
- music information retrieval.
Other research interests:
I am interested in many other areas, including:
but, rather unfortunately, there is not enough time for me to
actively pursue these interests at the present time.
- theoretical computer science (formal languages, automata, complexity, ...)
- parallel distributed processing,
- neurobiology and biocybernetics,
- meta-mathematics and formal Logic;