NetLiN – An Efficient and Effective Method to Prevent the Spread of Fuzziness in Decomposed Fuzzy Systems

W. Brockmann and O. Huwendiek (Germany)


Decomposed fuzzy systems, spread of fuzziness, NetLiN approach, low-cost applications


Complex fuzzy systems suffer from the knowledge engineering bottleneck which is caused by an exponential growth of the number of rules, called curse of dimensionality. Large single-staged fuzzy system thus get intractable with an increasing number of input variables. Therefore decomposition is used. Unfortunately the tractability is not increased in the same order as the curse of dimensionality is reduced. One reason is the spread of fuzziness. Several approaches have been introduced to tackle this problem. But each has some limitations. This paper discusses the reasons and describes an alternative which aims at good manageability of systems with many input variables and inexpensive implementation. It focuses especially on the knowledge engineering because manageability is very important for practical use, especially in control applications.

Important Links:

Go Back