Complexity reduction of a rational general form

This paper is motivated by the fact that application of B-spline and fuzzy techniques are popular engineering tools, however, their use is restricted by their exponential complexity. The main goal of this paper is to define a complexity reduction algorithm to the general form adopted in fuzzy logic and B-spline techniques. The key idea of this work is to store the huge number of parameters of the general form in a special compressed form offline, in a way that the decompression can then be executed by the same general form online, which enables us to regenerate not the whole parameter array, but only those parameters that are necessary to be considered. This idea is supported by the recently published fuzzy rule base reduction approach based on singular value decomposition. An example showing the effectiveness of the SVD reduction is presented.