The aim of this paper is to discuss the approximate rea- soning problems with interval-valued fuzzy environments based on the fully implicational idea. First, this paper constructs a class of interval-valued fuzzy imp...The aim of this paper is to discuss the approximate rea- soning problems with interval-valued fuzzy environments based on the fully implicational idea. First, this paper constructs a class of interval-valued fuzzy implications by means of a type of impli- cations and a parameter on the unit interval, then uses them to establish fully implicational reasoning methods for interval-valued fuzzy modus ponens （IFMP） and interval-valued fuzzy modus tel- lens （IFMT） problems. At the same time the reversibility properties of these methods are analyzed and the reversible conditions are given. It is shown that the existing unified forms of α-triple I （the abbreviation of triple implications） methods for FMP and FMT can be seen as the particular cases of our methods for IFMP and IFMT.展开更多
The weights of criteria are incompletely known and the criteria values are incomplete and uncertain or even default in some fuzzy multi-criteria decision-making problems.For those problems,an approach based on evident...The weights of criteria are incompletely known and the criteria values are incomplete and uncertain or even default in some fuzzy multi-criteria decision-making problems.For those problems,an approach based on evidential reasoning is proposed,in which the criteria values are integrated on the basis of analytical algorithm of evidential reasoning,and then nonlinear programming models of each alternative are developed with the incomplete information on weights.The genetic algorithm is employed to solve the models,producing the weights and the utility interval of each alternative,and the ranking of the whole set of alternatives can be attained.Finally,an example shows the effectiveness of the method.展开更多
Different fuzzy reasoning methods were gave by choosing different fuzzy counters.This article generally introduced the basic structure of fuzzy controller,and compared and analysised the reasoning effect of fuzzy reas...Different fuzzy reasoning methods were gave by choosing different fuzzy counters.This article generally introduced the basic structure of fuzzy controller,and compared and analysised the reasoning effect of fuzzy reasoning methods and the effect of computer simulating control basicly on different fuzzy counters.展开更多
Based on a presented inference algorithm of fuzzy reasoning, a fuzzy reasoning system is made up. A method of modeling the fuzzy reasoning system, and the setting up of the reasoning knowledge based and reasoning rule...Based on a presented inference algorithm of fuzzy reasoning, a fuzzy reasoning system is made up. A method of modeling the fuzzy reasoning system, and the setting up of the reasoning knowledge based and reasoning rules are studied in this paper. Then a heuristic inference algorithm is presented according to the system.展开更多
传统FCM算法对初值的依赖性过大且欧氏距离只适用于处理数值型及特征空间为超球结构的数据集。为此,利用模糊粗糙集思想,结合ReliefF技术,提出了一种基于模糊粗糙集的特征加权聚类算法(FRS-FCM),并将此算法应用到集成入侵检测中,通过有...传统FCM算法对初值的依赖性过大且欧氏距离只适用于处理数值型及特征空间为超球结构的数据集。为此,利用模糊粗糙集思想,结合ReliefF技术,提出了一种基于模糊粗糙集的特征加权聚类算法(FRS-FCM),并将此算法应用到集成入侵检测中,通过有效地聚类和集成学习来提高入侵检测的检测率,降低误检率,并较大地提高低频攻击的检测率。最后利用KDD Cup 99数据集进行的仿真实验验证了该方法的可行性与有效性。展开更多
基金supported by the National Natural Science Foundation of China(60774100)the Natural Science Foundation of Shandong Province of China(Y2007A15)
文摘The aim of this paper is to discuss the approximate rea- soning problems with interval-valued fuzzy environments based on the fully implicational idea. First, this paper constructs a class of interval-valued fuzzy implications by means of a type of impli- cations and a parameter on the unit interval, then uses them to establish fully implicational reasoning methods for interval-valued fuzzy modus ponens （IFMP） and interval-valued fuzzy modus tel- lens （IFMT） problems. At the same time the reversibility properties of these methods are analyzed and the reversible conditions are given. It is shown that the existing unified forms of α-triple I （the abbreviation of triple implications） methods for FMP and FMT can be seen as the particular cases of our methods for IFMP and IFMT.
基金supported by the National Natural Science Foundation of China(7077111570921001)and Key Project of National Natural Science Foundation of China(70631004)
文摘The weights of criteria are incompletely known and the criteria values are incomplete and uncertain or even default in some fuzzy multi-criteria decision-making problems.For those problems,an approach based on evidential reasoning is proposed,in which the criteria values are integrated on the basis of analytical algorithm of evidential reasoning,and then nonlinear programming models of each alternative are developed with the incomplete information on weights.The genetic algorithm is employed to solve the models,producing the weights and the utility interval of each alternative,and the ranking of the whole set of alternatives can be attained.Finally,an example shows the effectiveness of the method.
文摘Different fuzzy reasoning methods were gave by choosing different fuzzy counters.This article generally introduced the basic structure of fuzzy controller,and compared and analysised the reasoning effect of fuzzy reasoning methods and the effect of computer simulating control basicly on different fuzzy counters.
文摘Based on a presented inference algorithm of fuzzy reasoning, a fuzzy reasoning system is made up. A method of modeling the fuzzy reasoning system, and the setting up of the reasoning knowledge based and reasoning rules are studied in this paper. Then a heuristic inference algorithm is presented according to the system.
基金Supported by National Natural Science Foundation of China（61473176,61105077,61402260,61074149） the Excellent Young and Middle-Aged Scientist Award Grant of Shandong Province of China（BS2012DX026,BS2013DX043） the Open Program from theState Key Laboratory of Management and Control for Complex Systems（20140102）
文摘传统FCM算法对初值的依赖性过大且欧氏距离只适用于处理数值型及特征空间为超球结构的数据集。为此,利用模糊粗糙集思想,结合ReliefF技术,提出了一种基于模糊粗糙集的特征加权聚类算法(FRS-FCM),并将此算法应用到集成入侵检测中,通过有效地聚类和集成学习来提高入侵检测的检测率,降低误检率,并较大地提高低频攻击的检测率。最后利用KDD Cup 99数据集进行的仿真实验验证了该方法的可行性与有效性。
