Fuzzy reasoning is a method of uncertainty reasoning,it has become very popular recently in various fields such as network,control systems and artificial intelligence.Most of literature about reasoning methods are based on fuzzy set.Interval-valued fuzzy set is considerably easier to handle daily reasoning than fuzzy set in practice.This text mainly investigates interval-valued fuzzy reasoning methods.The main results and innovation points,obtained in this dissertation,may be summarized as follows:In the first chapter,we combine reverse triple I principle with intervalvalued fuzzy set,give the reverse triple I algorithms based on intervalvalued fuzzy inference.Also,the robustness of interval-valued reverse triple I algorithms are discussed.Results show that the robustness of intervalvalued fuzzy inference is directly linked to the selection of interval-valued fuzzy connectives.In the second chapter,we combine reverse triple I principle with intervalvalued Schweizer-Sklar t-norms,propose flexible interval-valued reverse triple I algorithms,and give the solutions of the new proposed algorithms.Also,the robustness of interval-valued Schweizer-Sklar reverse triple I algorithms are discussed.In the third chapter,the definition of similarity based on interval-valued is proposed,and the corresponding properties are proved.Also,the robustness of triple I method based on interval-valued is investigated from the perspective of similarity.In the forth chapter,as for the defects of CRI and triple I method,we combine quintuple implication algorithms with interval-valued fuzzy set,give the quintuple implication algorithms based on interval-valued fuzzy inference and the solutions of the new proposed algorithms.Also,the robustness of interval-valued quintuple implication algorithms are discussed. |