Computational Properties of Fuzzy Logic Deduction

. This paper studies computational properties of fuzzy logic deduction and compares them with the standard inference methods. The principles of deduction in fuzzy logic are explained and algorithms for its computer realization are described. Basic algorithm has exponential complexity with respect to the number of antecedent variables, and in case of fuzzy observations we are able to improve its performance only by constant factor. 1 Introduction In this contribution, we study computational and algorithmic aspects of fuzzy logic deduction. This type of fuzzy inference mechanism was proposed and investigated by V. Nov#k in, e.g. [6] and it has been demonstrated, that its properties make it suitable for applications in control systems as well as in decision making. Algorithmic aspects of the most commonly used Mamdani-Zadeh inference mechanism were extensively studied [5], where was shown that computational complexity grows exponentially with the number of antecedent variables. However, ..