关于模糊规则独立性问题的研究

基于模糊关系给出了模糊规则独立性的一种新的定义,讨论了模糊规则独立性的相关性质,从而提出了一种利用模糊规则独立性对规则库进行划分的方法,进而将规则库划分为独立规则和耦合规则两部分并建立了该规则库的推理公式;给出了Mamdani 组合推理方法与并组合独立推理方法之间的关系;最后,通过所建立的实际模糊系统给出了模糊规则库的具体划分,验证了这种划分的正确性.     

一种混合的模糊时间序列预测模型

随着理论与应用的需要,对模糊时间序列模型的研究和应用越来越深入。提出从论域的划分和模糊规则的提取两个方面对传统模型进行改进。模型首先采用自动聚类的方法对论域进行划分,并在此基础上建立具有权重的模糊规则;然后,利用粒子群算法对模型进行优化,进一步提高预测精度;最后,将Alabama大学入学人数作为本模型的实验数据。实验结果表明该模型是可行的,其预测结果明显优于参照预测模型。     

剩余格上的L滤子

剩余格为模糊逻辑和模糊推理提供了一种良好的代数结构,滤子是剩余格中一个十分重要的概念,它在基于剩余格的模糊逻辑代数语义的研究中,扮演着一个关键的角色。本文基于Pavelka所提出的广义MP规则和真值提升规则,研究基于这两种推理规则的演绎系统的代数化问题。引入L滤子的概念,讨论这些滤子之间的关系,并给出它们的一些代数刻画。     

集对分析在我国经济增长模糊规则提取中的应用

针对粗糙集和模糊聚类方法提取我国经济增长模糊规则算法复杂的问题,把集对分析用于我国31个省市经济增长模糊规则提取。结果显示,不仅算法简明,而且能同时提取宏观层次上的经济增长规则—固定资产投资对GDP的拉动效果要大于人力资源对GDP的拉动效果,而且还从微观层次上揭示出各省的经济增长规则,为我国十二五经济发展规划的实施提供决策参考。     

连续值域信息系统的规则提取与知识约简

研究连续值域信息系统的规则提取和知识约简，给出规则提取的完整方法和2种新的知识约简的概念，讨论各种协调集之间的关系，探讨系统允许误差变化对规则提取的影响。     

基于多维模糊关联规则的道路交通阻断分析方法

交通阻断成因复杂,与气象环境、道路线形、车辆状态以及交通环境等多因素相关。由于缺乏对造成交通阻断相关因素间潜在关联的研究,交通阻断管控一直是公路管理,特别是高速公路管理的难点。本文提出了一种基于多维模糊关联规则的道路交通阻断分析方法,发掘交通阻断的潜在规律和各因素间的关联关系。首先在国家现有相关划分体系和大量交通阻断(事件)案例的基础上,根据道路管理实际需求,建立了交通阻断多维属性模型,然后利用基于FCM的模糊关联规则,挖掘阻断因素的多维属性的依存关系,得到面向道路交通阻断分析的多维模糊关联规则。通过研究成果的实践应用,证明关联规则可以为道路交通阻断预防和管理提供有效支持,在道路交通阻断分析领域有着良好的应用前景。     

粗糙集方法在路口交通模糊控制中的应用

交通灯的控制是城市交通系统中一个重要的问题.对交通过程采用模糊控制,模糊规则的获取以及模糊规则的客观性问题一直是众多学者努力研究的问题.粗糙集理论的提出主要是用来处理不完整和不确定信息,可以用来观察、测试数据并进行逻辑推理.本文采用粗糙集的方法对模糊规则的生成进行处理,生成相对客观的模糊规则.并结合路口交通的模糊控制问题,阐述该方法的实用性以及生成的模糊规则的客观性.     

退化问题中Bland规则的改进

Bland规则是退化问题中避免循环的一种方法，本给出了Bland规则的两种改进形式，并从理论上证明了规则Ⅲ为有限规则，另外，还对一个退化问题给出了相应于不同规则的解法。     

一类模糊合作博弈资源与收益分配研究

本文的目的是提出一类具有特殊结构的模糊合作博弈的资源分配方法。具体地说,此类模糊合作博弈的特征函数具有Choquet积分的表示形式。我们提出的资源分配方法可以保证所有局中人支付之和最大,从而解决了模糊合作博弈中会形成什么样的模糊联盟结构的问题。针对改进了的模糊联盟结构,本文提出了一种分配规则。该规则可以实现局中人的个体理性与集体理性,从而保证了模糊联盟结构的稳定性。     

同一水平下模糊规则信息熵的描述

给出了模糊知识系统及模糊决策逻辑公式的定义,在此基础上描述了模糊决策逻辑公式及模糊知识系统下模糊规则的信息熵,讨论了模糊规则信息熵的相关性质;其次,利用模糊规则信息熵对模糊规则进行了分类、评价,从而为建立合理的模糊系统提供了一种有效的判定方法;最后,通过实例验证了所提出理论的正确性.     

Towards data driven selection of a penalty function for data driven Neyman tests

The data driven Neyman statistic consists of two elements: a score statistic in a finite dimensional submodel and a selection rule to determine the best fitted submodel. For instance, Schwarz BIC and Akaike AIC rules are often applied in such constructions. For moderate sample sizes AIC is sensitive in detecting complex models, while BIC works well for relatively simple structures. When the sample size is moderate, the choice of selection rule for determining a best fitted model from a number of models has a substantial influence on the power of the related data driven Neyman test. This paper proposes a new solution, in which the type of penalty (AIC or BIC) is chosen on the basis of the data. The resulting refined data driven test combines the advantages of these two selection rules.     

Discrimination with jointly equicorrelated multi-level multivariate data

In this article we study a linear as well as a quadratic discriminant function for multi-level multivariate repeated measurement data under the assumption of multivariate normality. We assume that the m-variate observations have jointly equicorrelated covariance structure in addition to a Kronecker product structure on the mean vector. The new discriminant functions are very effective in discriminating individuals when the number of observations is very small. The proposed classification rules are demonstrated on a real data set. The error rates of the proposed classification rules are found to be much less than the error rates of the traditional classification rules, when in fact the traditional classification rules fail most of the time owing to the small sample sizes.     

Using FSBT technique with Rough Set Theory for personal investment portfolio analysis

This study proposes a novel Forward Search and Backward Trace (FSBT) technique based on Rough Set Theory to improve data analysis and extend the scope of observations made from sample data to solve personal investment portfolio problems. Rough Set Theory mathematically classifies data into class sets. The class set with the most objects may generate one decision rule. The rules generated from RST are rough and fragmented, that are very difficult to interpret the information. An empirical case is used to generate more than 85 rules by the RST method in comparison with FSBT method which only generated 14 rules. This result can show our proposed method is better than traditional RST method based on class sets that contain the most objects. Much of human knowledge is described in natural language. It is a very important thing to convert information from computer databases into normal human language. Sample data taken from features with the same backgrounds are used to compile different portfolios that investment companies and investment advisors can employ to satisfy the investor’ needs. The method not only can provide decision-making rules, but also can offer alternative strategies for better data analysis. We believe that the FSBT technique can be fully applied in research on investment marketing.     

Multi-criteria classification – A new scheme for application of dominance-based decision rules

We are considering the problem of multi-criteria classification. In this problem, a set of “if … then …” decision rules is used as a preference model to classify objects evaluated by a set of criteria and regular attributes. Given a sample of classification examples, called learning data set, the rules are induced from dominance-based rough approximations of preference-ordered decision classes, according to the Variable Consistency Dominance-based Rough Set Approach (VC-DRSA). The main question to be answered in this paper is how to classify an object using decision rules in situation where it is covered by (i) no rule, (ii) exactly one rule, (iii) several rules. The proposed classification scheme can be applied to both, learning data set (to restore the classification known from examples) and testing data set (to predict classification of new objects). A hypothetical example from the area of telecommunications is used for illustration of the proposed classification method and for a comparison with some previous proposals.     

Hardware-in-the-loop experimental study on a fractional order networked control system testbed

Networked Control Systems (NCS) are of great interest in many industries because of their convenience in data sharing and manipulation remotely. However, there are several problems along with NCS itself due to the uncertainties in network communication. One issue inherent to NCS is the network-induced delays which may deteriorate the performance and may even cause instability of the system. Therefore a controller which can make the plant stable at large values of delay is always desirable in NCS systems. Our past work on Optimal Fractional Order Proportional Integral (OFOPI) controller showed that fractional order PI controllers have larger jitter margin (maximum value of delay for which system is stable) for lag-dominated systems when compared to traditional Proportional Integral Derivative (PID) controllers, whereas integer order PID controllers have larger jitter margin for delay-dominated systems. This paper aims at the design process of a tele-presence controller based on OFOPI tuning rules. To illustrate this, an extensive experimental study on the real-time Smart Wheel networked speed control system is performed using hardware-in-the-loop control. The real-time random delay in the world wide network is collected by pinging different locations, and is considered as the delay in our simulation and experimental systems. Comparisons are made with existing integer order PID controller. It is found that the proposed OFOPI controller is a promising controller and has faster response time than the traditional integer order PID controllers. Since the plant into consideration viz. the Smart Wheel is a delay-dominated system, it is verified that PID achieves larger jitter margin as compared to OFOPI tuning rules. Simulation results and real-time experiments showing comparisons between OFOPI and OPID tuning rules prove the significance of this method in NCS.     

基于控制图的马田系统马氏空间生成机理研究

马田系统是以马氏距离为测量尺度,通过选取正常样本构建马氏空间,对多元系统进行诊断和预测的分类技术。马氏距离对样本数据的变化非常敏感,因此用于构建马氏空间的正常样本的数据质量直接影响到分类的准确率。实际应用中正常样本的选取大多依据主观经验判断,缺乏客观规范的选择机制。本文提出基于控制图的马氏空间生成机理,先由专家选取的正常样本构建初始马氏空间,再以每个正常样品在初始马氏空间和对应的缩减马氏空间上的马氏距离增量作为新的测量尺度,以此建立单值控制图,利用控制图稳定性判定规则剔除异常数据,从而得到稳定状态的马氏空间。实验分析结果表明该方法的有效性且提高了马田系统分类的准确率。     

Lorenz comparisons of nine rules for the adjudication of conflicting claims

Consider the following nine rules for adjudicating conflicting claims: the proportional, constrained equal awards, constrained equal losses, Talmud, Piniles’, constrained egalitarian, adjusted proportional, random arrival, and minimal overlap rules. For each pair of rules in this list, we examine whether or not the two rules are Lorenz comparable. We allow the comparison to depend upon whether the amount to divide is larger or smaller than the half-sum of claims. In addition, we provide Lorenz-based characterizations of the constrained equal awards, constrained equal losses, Talmud, Piniles’, constrained egalitarian, and minimal overlap rules.     

考虑供应链影响时基于粗糙集方法的企业信贷风险评估

对于考虑供应链时的企业信贷风险评估问题,提出基于粗糙集的解决办法.首先,根据样本数据建立决策信息表;然后采用等间距法对决策信息表的连续属性值进行离散化,并且应用辨识矩阵求出最小约简;最后,应用启发式值约简算法求出决策规则.试验计算结果表明,所提出的方法对企业的信贷等级能够进行有效的预测.     

Stability of quadrature rule methods for first kind volterra integral equations

Gladwin [4] proved that Newton-Gregory formulas of order larger than 2 produce unstable algorithms when applied to nonlinear Volterra integral equations of the first kind. It is shown that similar results are true for all interpolatory quadrature rules using equidistant nodes. Upper bounds for the error order of quadrature rules, which lead to stable methods are given. Some higher order stable methods are indicated.     

Techniques for the assessment of influence in DEA

In this paper, we address the problem of assessing the influence of a given unit in the sample when evaluating efficiency by using both radial and nonradial DEA models. According to the values of some new measures we define, the efficient units exhibiting a higher deal of influence will be classified for a further checking. Then the analyst will have to decide whether they are contaminated by data errors or not. The interest of these techniques lies in that they provide rules to detect influential observations which may avoid data checking, which is often costly (particularly with large samples).