Approximating Fuzzy Control Strategies via CRI
نویسندگان
چکیده
We start from the observation that Zadeh’s compositional rule of inference (CRI) is a strategy to determine approximately a roughly given control function. From this point of view the problem to solve a system of fuzzy relation equations also becomes the problem to determine approximately such a control strategy. This gives a natural interpretation for approximate solutions of unsolvable systems of relation equations. Therefore we discuss and generalize some approaches and results about approximate solutions of systems of relation equations. Finally we discuss here how a choice of the t-norm which is involved in the sup-t-composition modifies the solvability behavior.
منابع مشابه
Pointwise Defined CRI-Based Aggregation Distributive Operators Are Trivial
In a rather general, essentially aggregation operator based discussion of the traditional fuzzy control strategies known as FATI and FITA strategies, a way to reduce these strategies to one another has been to define pairs of aggregation distributive aggregation operators. In this paper it is shown that for some often used special cases this reduction condition allows only the set theoretic uni...
متن کاملFuzzy Inference Method for Intelligent Artificial System
A fuzzy control system which is a typical system utilizing fuzzy model is mainly using the Max-Min CRI (Compositional Rule of Inference) method by Zadeh and Mamdani for fuzzy inference. But the Max-Min CRI method suffers from drawbacks including: error-prone weighting strategy, inefficient compositional rule of inference, and subjective formulation of membership functions. Because of these prob...
متن کاملOn the Suitability of the Bandler-Kohout Subproduct as an Inference Mechanism
Fuzzy Relational Inference (FRI) systems form an important part of approximate reasoning schemes using fuzzy sets. Especially, the compositional rule of inference (CRI) introduced by Zadeh has attracted the most attention so far. In this work we show that the FRI scheme based on the BandlerKohout subproduct along with a suitable realisation of the fuzzy rules, does possess all the important pro...
متن کاملFuzzy relation equations and approximate solutions. Some recent results
For theoretical fuzzy control it is a well known strategy to transform a system of control rules into a system of relation equations. Because these systems of relation equations are not always solvable, solvability criteria and approximate solutions have been discussed. We reconsider some of the results in this field and extend them using more recent results on t-norm based fuzzy logics and on ...
متن کامل