Logo of the Department

Department of Theoretical Computer Science and Mathematical Logic

Charles University, Faculty of Mathematics and Physics

Logo of the Faculty
 Home   People   Research   Teaching   Services   Contact 

Constraint Satisfaction and Optimisation Research Group

Contact: prof. RNDr. Roman Bartak, Ph.D.

The group performs basic and applied research in the areas of satisfiability and discrete optimisation problems by applying techniques such as constraint and Boolean satisfiability, (constraint) logic programming, and knowledge reasoning. We do research on general problem solving techniques including search and consistency algorithms (motivated by artificial intelligence), on algorithms for solving particular combinatorial optimisation problems such as scheduling, and on formal properties of problems such as problem complexity. Currently the main application areas of our research are planning and scheduling. We are working on generally applicable planning and scheduling engines, on formal specification of problems (problem modelling), and on automated extraction of knowledge from models that is useful for problem solving. Our current focus is on techniques integrating temporal, resource, and logical reasoning.


Interconnection networks

Contact: Mgr. Petr Gregor, Ph.D.

We study theoretical aspects of interconnection networks for parallel and distributed computing. In particular, we are interested in structural properties of hypercubic networks, design of algorithms, and the computational complexity of related problems. Futhermore, we explore applications of hypercubes in other areas such as data compression, coding theory, and extremal combinatorics.

Recognizing, Analysis and Translations of Languages and Picture Languages

Contact: Martin Plátek, CSc.

Typical kind of theoretical results is an extension of the known taxonomies of language classes and of the computability and complexity results that concern the problems related to flexible automata or grammars and their constrains. In the main stream of our efforts we focus on the measures of the word-order freedom that is studied both by grammars and by automata. The degrees of nondeterminism is also studied. Several applications in computational linguistic of methods studied in this group are considered and developed. Namely the methods of modeling of analysis by reduction and its relationship to the (extended) dependency syntax is studied. Moreover it seems to be important to develop new methods for automatic restarting automata construction. One of the ways to do that is the use of machine learning methods based e.g. on genetic algorithms. We extend our effort also to the study of two-dimensional languages.


Smart Systems Group

Contact: doc. RNDr. Iveta Mrázova, CSc.

Smart systems for the new millennium need to meet the challenge of flexibility and customized design requirements. From this point of view, one of the most decisive issues represents the design, analysis, and application of effective machine learning algorithms and data mining tools based in particular on artificial neural networks, fuzzy logic, evolutionary programming, and automata theory. The areas of our research include:

  • specification of the properties characteristic for important features, significant input patterns and those tendencies in the system development, which might indicate its future change
  • development of new adaptive methods and formal tools suitable for a fast and reliable detection and identification of significant data patterns
  • design and verification of methods for optimization of the applied knowledge extraction techniques (neural networks and other data mining tools)
  • exact analysis of the proposed methods and tools and their applications
  • interpretation, verification, and visualization of the knowledge extracted from the analyzed data