MIT nyitóoldal

   Complete list

Application of hybrid intelligent methods in modeling

OTKA project
2000 - 2003

Local supervisor: Pataki Béla
Official project supervisor: BME MIT

In the tasks of modeling complex systems it is typical that mathematical models are not applicable or such models have only limited value. Due to the complexity and the hardly measurable parameters of the physical equations characterizing such systems the precise physical model is hard to construct. In several such cases there is a possibility to make qualitative models based partly on physical laws partly on the heuristic knowledge of the professionals. Beyond these two sources the collected empirical data could be used in the modeling task.
Most of the research carried out in the last few years captured only one of the aspects of the previously mentioned ones. Neural models used the empirical measured data, fuzzy models applied the qualitative knowledge of the professionals, etc. Because in most of the situations all the mentioned knowledge types are more or less available it is reasonable to use hybrid models in such tasks. During this project the main goals are to investigate the cooperation of the subsystems based on different modeling paradigm, and to work out multilevel methods mixing traditional and new modeling principles.

Further information about the project:
Department homepage:
Official email address: