基于云模型的区间直觉模糊数多属性群决策

针对属性值为区间直觉模糊数的多属性群决策问题,考虑到模糊性和随机性对群决策过程及结果的影响,本研究将利用云模型理论结合区间直觉模糊数的特征,运用灰色关联系数法和信息熵理论确定专家和属性权重,通过信息集结构建综合评价云模型.不同于传统的区间直觉模糊数的排序方法,本研究利用云模型的3En规则将区间直觉模糊数进行云转换并通过云相似度确定方案的综合评价值和犹豫度,然后对决策方案进行比较分析.研究结果表明:该方法能够科学有效地进行决策,进而为决策方提供科学依据.     

考虑碳限额的制造/再制造混合系统生产优化决策

针对属性权重部分未知且专家权重完全未知的多粒度语言大群体决策问题,提出一种基于云模型的决策方法.首先,构建一种基于信任关系的专家权重求解模型来计算专家权重;其次,将多粒度语言转换为云模型并进行聚类;然后,构建一致性优化模型来求解属性权重,从而得到各个方案的综合评价值并对方案进行排序.所构造的专家赋权模型可以有效解决大群体决策过程中决策人数众多、无法客观给出专家权重信息的问题,而且通过定义的直觉信任函数,还可以对专家之间的信任关系进行刻画,充分挖掘专家之间的信息;将多粒度语言转换为云模型,可以有效刻画语言信息的模糊性和随机性,从而避免信息的丢失和失真.     

群体分类偏好下的双重语言信息融合聚类方法

研究一类基于双重信息融合的群体聚类方法。依据偏好信息下专家意见相似关系挖掘群体分类偏好信息;考虑专家决策依据向量的相似程度,设计一致性和非一致性测度指标,以表征双重维度下群体聚类的一致及差异度;以群体聚类结果差异最小为目标构建规划模型,测算属性权重,并以编网聚类的思想给出聚类结果。算例研究验证了所提出方法的科学性和合理性。     

A multistage risk decision making method for normal cloud model considering behavior characteristics

It is difficult to make scientific decisions for multistage risk decision-making problems, which are mainly influenced by these challenges, such as uncertainty of information, the passage of time and complexity of the problems The multistage risk decision-making problem is analyzed for normal cloud model considering behavior characteristics in this paper. Firstly, considering the fuzziness and randomness of multi-stage risk decision-making information, the relative entropy and modified relative entropy are defined as the distance measures for the normal cloud model from the perspective of full utilization of information. Secondly, the prospect theory is extended to normal cloud model environment. Considering the behavior characteristics of reference dependence, two reference points are set up considering the short-term development and long-term development. Thirdly, the optimization model is built to obtain the attribute weights and stage weights based on the principle of maximizing the difference between ranking values. The ranking of alternatives is determined by the final ranking values. The case analysis and the method comparison illustrate the feasibility and rationality of the proposed method.     

基于云模型的Pythagorean模糊语言多属性决策方法

针对专家给出的属性值为Pythagorean模糊语言且专家权重与属性权重均未知的多属性决策问题进行了研究,提出一种基于云模型的多属性决策方法。首先,根据Pythagorean模糊语言决策信息的距离熵计算得到属性权重;其次,计算决策矩阵间的距离从而得到各决策专家权重;再次,构建Pythagorean模糊云模型决策矩阵并利用专家权重和属性权重进行信息集结;最后,基于TOPSIS方法求取正、负理想解,依据理想解计算各方案贴近度并据此对各备选方案进行排序选择。案例分析表明,该方法优化了复杂环境下的决策,避免了决策信息的丢失,能够较好解决决策信息的不确定性和决策过程的随机性,具有一定的可行性和有效性。     

基于对称交互熵的犹豫模糊信息相似度及聚类应用

研究面向犹豫模糊信息的聚类方法。首先,定义犹豫模糊相对熵、对称交互熵,并基于信息论的角度提出一个新的犹豫模糊相似度公式;然后,利用相似度公式构造相似系数矩阵,基于编网聚类方法对犹豫模糊集进行聚类;最后,通过算例验证了所提出方法的有效性。     

基于云模型的改进FMEA风险评估方法

针对风险评估过程中的模糊性和随机性两个不确定性因素,采用云模型对语义评价变量进行量化,并通过云的合成对专家的评价结果进行集结,构建云风险评价矩阵。将风险因子O、S、D期望重要度与云的相似度相结合计算风险因子的权重。考虑到传统故障模式与影响分析（Failure Mode and Effects Analysis,FMEA）中故障模式风险值排序的缺陷,基于云的距离测度算法提出云-VIKOR的风险排序方法。最后以实例验证了所提方法的精确性和有效性。     

基于云模型的不确定性大群体多属性决策方法

针对传统不确定性大群体多属性决策方法中只考虑决策信息的模糊性,没有考虑信息的随机性这一问题,提出了一种基于云模型的多属性决策方法,从而用于解决由多个小群体组成的不确定性大群体决策问题。首先将不确定语言评价值转化为一维正态云;其次采用决策者主观确定和一致性分析相结合的方法确定针对不同决策对象的小群体权重,进而生成综合云;然后提出了一种改进的云相似度算法作为云模型距离的度量,通过比较各方案综合云与最优云的相似度对方案排序。最后通过实例验证了所提方法的可行性和有效性。     

An objective and interactive-information-based feedback mechanism for the consensus-reaching process considering a non-support degree for minority opinions

The consensus-reaching process (CRP) to achieve higher unanimity and ensure common agreement before deriving a final decision has become an important procedure in group decision-making problems. The demand for high-quality decision results has motivated the development of large-scale group decision-making (LGDM). In such cases, the issue of minority opinion has gained awareness due to the related effects on enhancing consensus and decision quality. A minority opinion cannot exert an effect unless the majority attach importance to whether that opinion is supported or not. To reflect the effect of minority opinions on consensus, this paper establishes a LGDM framework with an objective and interactive-information-based feedback mechanism for the CRP. Given the natural forms of human expression, multi-granular linguistic information and a 2-tuple linguistic model are used. First, initial weights are objectively assigned to decision-makers (DMs) to weaken the impact of the majority. Subsequently, a non-support degree function is newly defined to reflect the extent to which other DMs dissent from a minority opinion. More importantly, feedback rules are constructed to make corresponding adjustments to the powers of discourse among all DMs in the attempt to reach consensus. Finally, the proposed three-phase LGDM framework is applied to new product development (NPD), and simulation experiments are conducted based on two algorithms to verify the framework's applicability and feasibility.     

A method for large group decision-making based on evaluation information provided by participators from multiple groups

Large group decision-making (LGDM) is a special group decision-making (GDM) problem, in which a large number of persons take part in decision process, while research concerning this issue is still relatively scarce. The objective of this paper is to develop a method to solve the LGDM problem, in which a large number of persons from multiple groups take part in the decision process and express their personal evaluations on the alternatives according to the pre-established identifier set. In the method, the percentage distribution on evaluations of each group concerning each alternative is determined. The decision weight of each group concerning each alternative is obtained by aggregating the subjective weight, which is provided by the organizer, and the objective weight, determined according to the level of consensus among participators' evaluations. According to the percentage distributions and decision weights, the dominance degrees on pairwise comparisons of alternatives are calculated, and a ranking of alternatives can be determined using the PROMETHEE II method. Finally, an example is given to illustrate the use of the proposed method.     

基于概率语言术语集的云服务选择

为了解决云服务评估决策中QoS(Quality of Service,服务质量)属性的动态性刻画不足以及传统决策方法中用户主观因素过强的缺点,提出了一种基于概率语言术语集(Probabilistic Linguistic Term Set,PLTS)的选择方法.通过相似性权重与可靠性权重结合获取推荐权重,加入决策矩阵中...     

基于改进向量相似度的区间数动态多指标决策模型及应用

针对决策信息为区间数的不确定性动态决策问题,在属性权重和时间权重未知的情况下,基于改进向量相似度的方法,构建一种兼顾决策信息和决策偏好的动态多指标决策模型.利用区间型决策信息的相对相似性和属性重要度,构造相对相似度最小规划模型以确定指标权重;在综合考虑决策信息时间价值、决策者偏好的基础上,构建极大熵模型以确定时间权重;结合向量相似度计算存在的缺陷,提出一种基于向量投影思想的向量综合相似度测度方法,从而建立不确性动态决策模型,并通过实例分析检验该模型的合理性和有效性.     

基于加权多属性云的服务信任评估方法

针对云计算环境中服务信任的随机性和模糊性以及现有基于云模型的信任评估方法对时效性和推荐信任考虑不足的问题,提出一种基于加权多属性云的服务信任评估方法。首先,引入时间衰减因子为每次服务评价赋权重,从服务的多个属性细化信任评估粒度,通过加权属性信任云逆向生成器得到直接信任云;然后,根据评价相似度确定推荐实体的推荐权重,并计算得到推荐信任云;最后综合直接信任云和推荐信任云生成综合信任云,通过云相似度计算确定服务的信任等级。仿真结果表明,所提方法明显提高了服务交互成功率并有效抑制恶意推荐,能够更加真实地反映云计算环境中服务信任情况。     

云模型与用户聚类的个性化推荐

针对传统推荐系统数据稀疏、相似性计算方法导致共同评分用户少的问题,提出利用云模型定性概念与定量数值转换的优势,研究云模型、用户聚类的个性化推荐改进算法。用户对项目属性评价的偏好,转换为用户对加权综合云模型表示的数字特征的偏好。利用改进的聚类算法,对评分数据、原始用户属性标准化后的信息进行聚类;同时考虑用户兴趣的变化,结合用户之间项目属性评价的综合云模型的相似度、用户对项目评分的聚类、用户属性聚类这三种方法产生的邻居用户的并集进行推荐。理论分析和实验结果表明,提出的改进算法不但解决数据稀疏性带来的共同评分用户少的弊端,即使是在新用户的情况下,仍能获得较低的平均绝对误差和平均平方误差     

基于语言偏好序的多属性群决策方法及其在废旧物资循环利用中的应用

循环经济作为应对资源短缺威胁和绿色低碳转型需求的新型经济发展模式,正受到越来越广泛的关注.为了推动这一计划的实践进程,众多举措被提出并付诸行动,废旧物资循环利用体系建设是其中之一.考虑到与此类问题相关的决策过程中通常涉及多方面的制约因素且需要平衡各方意见,提出一种基于语言偏好序和拓展TOPSIS的多属性群决策方法.首先,利用数据挖掘和自然语言处理等技术,对相关新闻网站进行爬虫并获得大量的公共数据,再通过关键词提取和聚类确定决策属性及其权重;然后,采用基于多准则相互评估矩阵的专家互评方式,对参与决策的专家完成赋权;接着,运用语言偏好序、拓展的TOPSIS方法以及最小-最大优化模型依次实现决策信息表征、个体排序和集体排序的完整决策过程;最后,为了验证所提出方法的适用性,将其应用于废旧物资回收站点设置方案选择的实例,并进一步通过比较分析对其优势和特点展开具体说明.     

Managing information measures for hesitant fuzzy linguistic term sets and their applications in designing clustering algorithms

Recently, the Hesitant Fuzzy Linguistic Term Sets (HFLTSs) have been widely used to address cognitive complex linguistic information because of its advantage in representing vagueness and hesitation in qualitative decision-making process. Information measures, including distance measure, similarity measure, entropy measure, inclusion measure and correlation measure, are used to characterize the relationships between linguistic elements. Many decision-making theories are based on information measures. Up to now, distance, similarity, entropy and correlation measures have been proposed by scholars but there is no paper focuses on inclusion measure. This paper dedicates to filling this gap and the inclusion measure between HFLTSs are proposed. We discuss the relationships among distance, similarity, inclusion and entropy measures of HFLTSs. Given that clustering algorithm is an important application of information measures but there are few papers related to clustering algorithm based on information measures in the environment of HFLTS, in this paper, we propose two clustering algorithms based on correlation measure and distance measure, respectively. After that, a case study concerning water resource bearing capacity is illustrated to verify the applicability of the proposed clustering algorithms.     

网络安全风险评估的云决策

为了更合理地评估网络安全风险,利用云模型集成随机性和模糊性的优点,提出一种基于云模型的网络安全风险评估和决策方法。首先,通过采样系统正常状态信息,构造标准概念云;在进行风险评估时,采样处于风险状态时的信息,计算其云数字特征;然后利用改进的基于云滴距离的云相似度算法,计算与标准概念云的相似度,相似度最大的即为最终输出结果。最后,通过Kddcup99数据集进行模拟攻击及性能采样仿真实验。结果表明,该方法最大限度地保留了风险评估过程中固有的不确定性和模糊性,提高了评估结果的可信性。     

A multi-agent affective interactive MAGDM approach and its applications

Traditional multi-attribute group decision-making (MAGDM) methods focus on weights calculation of sub-attributes and experts' preferences, but lack the discussion on the decision-makers' affective interaction, and its influence on the decision preference and group consistency. To address this problem, the present study proposed a new multilayer affective computing model based on “personality–mood–emotion” pattern, under the multi-agent decision system framework. In addition, we introduced the group trending index and affection-preference incentive mechanism, which can help simulate MAGDM process and learn group experts' decision preferences. Further, we proposed a new multi-agent affective interactive MAGDM (MAAI-MAGDM) method, where we defined a novel group convergence index and an alternative decision entropy to explain the convergence process of decision and group consistency. Compared to the traditional MAGDM approaches, the proposed MAAI-MAGDM method fully considered the affective features of each expert, reduced the dependence on aggregation operators and weight analysis, alleviated the workload of group experts, and effectively reduced the complexity of decision-making calculation process. Finally, we verified that the proposed method can effectively assist the decision-making processes by employing two numerical cases.     

Services recommended trust algorithm based on cloud model attributes weighted clustering

There are many different cloud services available, each with different offerings and standards of quality. Choosing a credible and reliable service has become a key issue. To address the shortcomings of existing evaluation methods, we propose a service clustering method based on weighted cloud model attributes. We calculate user-rating similarity with the weighted Pearson correlation coefficient method based on service clustering, and then compute user similarity combined with the user service selection index weight. This method allows us to determine the nearest neighbors. Finally, we obtain the recommended trust of the service for the target user through the recommendation trust algorithm. Simulation results show that the proposed algorithm can more accurately calculate service recommended trust. This method meets the demand of users in terms of service trust, and it improves the success rate of user service selection.