Janis GRUNDSPENKIS

Professor Janis GRUNDSPENKIS

Head of the Department of Systems Theory and Design
Director of the Institute of Applied Computer Systems
Dean of the Faculty of Computer Science and Computer Systems
Riga Technical University
Kalku iela 1
Riga, LV 1658
Latvia

Phone: +371 7089506
Fax: +371 7089584
E-mail: [email protected]

Born: June 19, 1942, Riga, Latvia

 

Interests:

 

Brief Description of Main Research:

The research was focused on the development of a new approach to the modelling of complex technical systems with physically heterogeneous elements. The essence of this approach, called structural modelling, is automation of knowledge base development. Structural modelling is a systematic, partly formal and model-based method of knowledge acquisition. The two level concept of knowledge base has been proposed. The knowledge base is divided into two parts- a topological knowledge base and a deep knowledge rule base. The knowledge base supports several kinds of reasoning. The topological knowledge base consists of three structural models - a model of morphological structure and two kinds of models of functional structure. The topological knowledge base supports reasoning in logic. This allows to make decisions about relationships between elements and about functioning of a system in normal conditions. The deep knowledge rule base supports diagnostic and predictive reasoning if backward or forward chaining is executed, respectively. Hierarchy of frames has been chosen for knowledge representation. Methods and algorithms for analysis of various kinds of structures have been developed. Structural modelling approach has been implemented in automated structural modelling system ASMOS. Applications of structural modelling may be applied in at least two problem domains - at early stages of design and in technical diagnostics. In these problem domains ASMOS was used for structural modelling of several kinds of complex technical systems.

The ongoing research is focused on integration of deep knowledge rule base with an expert system shell. This will improve explanation facilities of prognosis expert systems. Another research project deals with the adding of the temporal dimension to the casual domain models. The temporal dimension will allow to use hybrid knowledge representation schemes. Hybrid knowledge-based systems will allow to produce an early prognosis (detection, isolation and identification) of faults while they are incipient.

Languages: Latvian, English, Russian

Education:

Experience:

Honours and Awards:

Professional Activities and Memberships:

Courses:

Recent/Representative Publications:

Research Projects:

 

I am a little bit conservative. The best evidence of this is that I have lived only in Riga and I have had only one employer - Faculty of Computer Science and Computer Engineering, Riga Technical University. I love my country and my family - my spouse, my daughter and my two sons. My hobbies are: tennis, alpine skiing and traditional jazz.

 



Last update 22.08.2005