On mathematical structures of fuzzy rough set algebras

In rough set theory, the lower and upper approximation operators are important notions defined by a binary relation. In this paper, we introduce a general type of relationbased fuzzy rough model determined by a triangular norm. Properties of fuzzy rough approximation operators are examined. The fuzzy rough approximation operators are also characterized by axioms. A comparative study of the fuzzy rough set algebra with other mathematical structures such as fuzzy topological spaces, fuzzy measurable spaces, and fuzzy belief structures is investigated.     

Short-term load forecasting based on fuzzy neural network

The fuzzy neural network is applied to the short-term load forecasting. The fuzzy rules and fuzzy membership functions of the network are obtained through fuzzy neural network learming. Three inference algorithms, i.e. the multiplicative inference, the maximum inference and the minimum inference, are used for comparison. The learning algorithms corresponding to the inference methods are derived from back-propagation algorithm. To validate the fuzzy neural network model, the network is used to Predict short-term load by compaing the network output against the real load data from a local power system supplying electricity to a large steel manufacturer. The experimental results are satisfactory.     

Analytical Structures of Typical Takagi - Sugeno Fuzzy PID Gain Scheduling Controllers

The analytical structure of a class of typical Takagi Sugeno (TS) fuzzy controllers is revealed in this paper.The TS fuzzy controllers consist of three or more trapezoidal input fuzzy sets, Zadeh fuzzy logic AND operator,fuzzy rules with linear consequent, and the centriod defuzzifier. The TS fuzzy controllers are proved to be accurately nonlinear PID controllers with gains continuously changing with process output. The analytical expressions of the variable gains of the TS fuzzy controllers are derived and their mathematical characteristics including the bounds and geometrical shape of the gain variation are analyzed. The resulting explicit structures show that the TS fuzzy controllers are inherently nonlinear PID gain scheduling controllers with variable gains in different regions of input space.     

Reliable fuzzy control with domain guaranteed cost for fuzzy systems with actuator failures

The reliable fuzzy control with guaranteed cost for T-S fuzzy systems with actuator failure is proposed in this paper. The cost function is a quadratic function with failure input. When the initial state of such systems is known, a design method of the reliable fuzzy controller with reliable guaranteed cost is presented, and the formula of the guaranteed cost is established. When the initial state of such systems is unknown but belongs to a known bounded closed domain, a notion of the reliable domain guaranteed cost (RDGC) for such systems is proposed. For two classes of initial state domain, polygon domain and ellipsoid domain, some design methods for reliable fuzzy controllers with the RDGC are provided. The efficiency of our design methods is finally verified by numerical design and simulation on the Rossler chaotic system.     

Intelligent Control on Hot Strip Coiling Temperature

A new intelligent control scheme for hot strip coiling temperature is presented. In this scheme, the prediction model of finishing temperature and the presetting model of main cooling zone are establish based on BP neural network, the feed-forward open-loop control model of main cooling zone is constructed based on T-S fuzzy neural network, a new improved structure of T-S fuzzy neural network is developed, and the feedback close-loop control model of precision cooling zone is obtained based on fuzzy control. The effectiveness of the proposed scheme has been demonstrated by computer simulation with a satisfactory result.     

A dynamic fuzzy clustering method based on genetic algorithm

A dynamic fuzzy clustering method is presented based on the genetic algorithm. By calculating the fuzzy dissimilarity between samples the essential associations among samples are modeled factually. The fuzzy dissimilarity between two samples is mapped into their Euclidean distance, that is, the high dimensional samples are mapped into the two-dimensional plane. The mapping is optimized globally by the genetic algorithm, which adjusts the coordinates of each sample, and thus the Euclidean distance, to approximate to the fuzzy dissimilarity between samples gradually. A key advantage of the proposed method is that the clustering is independent of the space distribution of input samples, which improves the flexibility and visualization. This method possesses characteristics of a faster convergence rate and more exact clustering than some typical clustering algorithms. Simulated experiments show the feasibility and availability of the proposed method.     

Intuitionistic Fuzzy Rough Sets Based on Two Intuitionistic Fuzzy Implicators

<正>This paper presents a general framework for the study of relation-based intuitionistic fuzzy rough sets determined by two intuitionistic fuzzy implicators.By employing two intuitionistic fuzzy implicators I and J,I -lower and J-upper approximations of intuitionistic fuzzy sets with respect to an intuitionistic fuzzy approximation space are first defined.Properties of(I,J) -intuitionistic fuzzy rough approximation operators are then examined.The connections between special types of intuitionistic fuzzy relations and properties of (I,J)-intuitionistic fuzzy approximation operators are also established.     

Marginal linearization method in modeling on fuzzy control systems

Marginal linearization method in modeling on fuzzy control systems is proposed, which is to deal with the nonlinear model with variable coefficients. The method can turn a nonlinear model with variable coefficients into a linear model with variable coefficients in the way that the membership functions of the fuzzy sets in fuzzy partitions of the universes are changed from triangle waves into rectangle waves. However, the linearization models are incomplete in their forms because of their lacking some items. For solving this problem, joint approximation by using linear models is introduced. The simulation results show that marginal linearization models are of higher approximation precision than their original nonlinear models.     

Takagi - Sugeno Fuzzy PID Gain - Scheduling Controllers

This paper aims at Takagi - Sugeno (TS) fuzzy controllers as gain scheduling (GS) schemes of PID controllers. A TS fuzzy controller employs arbitrary input fuzzy sets, product or Zadeh fuzzy logic AND, TS fuzzy rules with linear consequent, and the generalized defuzzifler containing the popular centrold defuzzifler as a special case. We first establish the relationship between the TS fuzzy controller and the linear PID controller. The TS ftizzy controller is accurately a nonlinear PID controller with gains continuously changing with Its process output. Then we point out that the TS fuzzy controller is closely related to the traditional gain scheduler. The gains of the TS ftizzy controller are determined by three two - Input - one - output fuzzy systems with singleton output fuzzy sets. Finally, as a demonstration, a simple TS fuzzy controller employing two linear input fuzzy sets, Zadeh fuzzy logic AND, and the popular centrold defuzzifler is designed to be the gain scheduler for the PID controller. Com     

Fuzzy optimization of space frame

Fuzzy concepts are introduced into structural optimization to solve fuzzy optimization problems with a crisp objective function and fuzzy constraints, also a non-membership function is used to convert fuzzy constrains into crisp constrains. Two models are discussed where the objective function considered is the volume of space frame and the fuzzy constrains are design limits by the axial strength, slenderness, deflection, thickness and diameter of space frame member.     

由伴随对构造的模糊控制器的控制性能

考虑了分别由9个模糊蕴涵算子的伴随对构造的模糊控制器．以直流串联电机为实例．验证了每个模糊控制器的控制品质，其控制品质是通过比较均方根差的大小来衡量的．与其他文献的结果相比，采用所提供的模糊蕴涵算子构造的模糊控制器，其控制品质更好．设计者也可以参考均方根差，选用合适的模糊蕴涵算子设计模糊控制器．     

一类模糊控制器的设计算法及其分类

给出了由模糊蕴涵算子构造模糊控制器的一般算法.根据模糊控制器的响应函数的拟合性质将其分为4种类型:具有泛逼近性的模糊控制器、只具有拟合性的模糊控制器、只具有阶跃响应性的模糊控制器和不定型的模糊控制器.     

由三I或CRI方法构造的模糊系统及其响应函数

文章介绍在采用单点模糊器和重心法解模糊器的情况下,分别基于三I方法和CRI方法,研究了由11个蕴涵算子构成的模糊系统及其响应函数。在基于三I方法的11个系统中,9个输出近似为阶跃函数,1个输出近似为拟合函数,1个具有函数逼近的泛性,从而后2个系统可用;基于CRI方法的11个系统中,7个输出近似为阶跃函数,2个输出近似为拟合函数,2个具有函数逼近的泛性,从而后4个系统可用。可见,从模糊系统的角度,三I方法效果不理想,而CRI方法基本可以接受,且相对较优。     

几个模糊控制器的控制性能

由模糊蕴涵算子构造了一个模糊控制器.首先计算出采用该模糊控制器控制直流串联电机的运行,当电机处于稳定状态时的均方根差;然后与一些文献中给出的具有较好控制品质的模糊控制器加以比较.结果表明,采用本文所构造的模糊控制器控制直流串联电机的运行,其控制效果会更好.     

直觉模糊逻辑算子的研究

给出了直觉模糊逻辑“补”、“与”、“或”及“蕴涵”算子的定义，并利用区间值模糊集与直觉模糊集之间的关系，给出了利用经典的模糊逻辑算子构造直觉模糊逻辑算子的三个定理。从而得到了构造直觉逻辑算子的新方法，这种方法无需验证其运算的封闭性，因而简单易行。文中用此方法构造出了一系列新的直觉模糊逻辑“补”、“与”、“或”及“蕴函”算子，将K.Atanassov最早提出的直觉逻辑模糊逻辑算子推广到了更一般情形。     

模糊统计法与蕴涵算子

对模糊统计法作了推广，得到了模糊蕴涵算子应满足的性质，并据此研究了十六种常用的模糊蕴涵算子，进而选出四种定义合理的算子．以这四种蕴涵算子为例，根据模糊统计的思想，阐述了如何确定每种蕴涵算子适用的场合．     

伴随对及其构造

先讨论伴随对定义及其等价条件,给出伴随对存在的一个充分条件;其次讨论40个模糊蕴涵算子和40个模糊圈乘算子的性质,并给出由模糊蕴涵算子构造模糊圈乘算子的方法;最后讨论40个算子对构成伴随对的情况.     

蕴涵算子族及其构造

给出了蕴涵算子族的定义,将蕴涵算子的性质进一步推广,给出了蕴涵算子族的16条常用性质.基于含参数的三角模和余三角模构造了42种蕴涵算子族.验证了所构造的蕴涵算子族是否满足16条常用性质,揭示了它们与几个重要蕴涵算子族之间的关系.     

关于模糊控制器性能与合成算子的研究

从改变合成算子角度出发,对Ｍａｍｄａｎｉ模糊控制器进行了改进,得到了一组可排序的模糊控制器,并对它们的性能进行了比较,获得了满意的结果。     

直觉区间值模糊逻辑算子

首先提出了用直觉区间值去表示命题的真值;然后定义了直觉区间值模糊逻辑上的算子－补,t-范、t－余范和蕴函算子,并讨论了它们与I[0,1]及[0,1]上的相应算子的内在联系,指出它们可用I[0,1]及[０,１]上相应逻辑算子表示。