Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

A new fuzzy-based multi-objective optimization concept for process control systems

: Bernard, T.; Sajidman, M.; Kuntze, H.-B.


Reusch, B.:
Computational intelligence. Theory and applications : International Conference. Proceedings
Berlin: Springer, 2001 (Lecture Notes in Computer Science 2206)
ISBN: 3-540-42732-5
ISSN: 0302-9743
Fuzzy Days <7, 2001, Dortmund>
International Conference on Computational Intelligence <7, 2001, Dortmund>
Fraunhofer IITB ( IOSB) ()

Control system design problems are generally multi-objective since they require several, generally conflicting requirements to be simultaneously met. In order to consider all individual objectives different vectorized solution concepts have been proposed in the past in which the problem is expressed as a set of algebraic inequalities. In this paper a new concept based on the theory of Fuzzy Decision Making will be introduced. According to the natural blurred character of the objectives and the human way of evaluating control performance the individual target functions will be expressed by Fuzzy membership functions which are aggregated to a common performance index by logical AND connection. The proposed Fuzzy-concept has been successfully applied to the control design of a complex industrial batch process which is characterized by heavy stochastic disturbances and large sensor time delay. By the employment of the developed optimization concept the implementing time of new controllers has been reduced and the product quality has been considerably improved.