Simplification and reduction of fuzzy rules

From MaRDI portal





This study is concerned with an important issue of complexity reduction of fuzzy rulebases. Given a rule-based system consisting of rules \[ \text{-- if }x_1\text{ is }A_{i1}\text{ and }\dots\text{ and }x_n\text{ is }A_{in}\text{ then }g={\mathbf a}^T_i{\mathbf x}+ b, \] \(i= 1,2,\dots, N\), where \({\mathbf x}= [x_1,x_2,\dots,x_n]^T\), the authors discuss various ways of reduction of the rulebase. Those include: (a) Similarity-driven simplification. This approach focuses on reduction of the rulebase by identifying redundant information that manifests in the form of similar fuzzy sets in the condition parts of the rules. Such fuzzy sets are then merged; in this way their number is reduced, and subsequently the number of rules goes down. (b) Rule reduction with the use of singular vector decomposition and orthogonal transforms. The performance of the methods is illustrated with the aid of some numerical benchmark problems. For the entire collection see [Zbl 1015.00012].











This page was built for publication: Simplification and reduction of fuzzy rules

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4437862)