分享兴趣，传播快乐，增长见闻，留下美好。
亲爱的您，
这里是LearningYard学苑！
今天小编给大家带来期刊论文精读系列四，
欢迎您的用心访问！
本期推文阅读时长大约6分钟，请您耐心阅读。
Share interest, spread happiness, increase knowledge, and leave beautiful.
Dear you,
This is the LearningYard Academy!
Today,the editor brings you the fourth series of intensive reading of journal papers,
Welcome your visit!
This tweet usually takes about 6 minutes to read. Please be patient and read.
本期小编为大家带来精读期刊论文《基于有限置信度和社会网络的大规模群体决策的共识达成过程》第四节和第五节内容，第四节给出了一个数值算例和比较分析来验证我们所提出的方法的优点。第五节提供了仿真分析和讨论，以说明所提出的方法的有效性。
In this issue, the editor brings you the content of sections 4 and 5 of the intensive reading journal paper "Consensus Reaching Process of Large-scale Group Decision-Making Based on Limited Confidence and Social Networks". Section 4 gives a numerical example and comparison analysis to verify the advantages of our proposed method. Section V provides simulation analysis and discussion to illustrate the effectiveness of the proposed method.
4. 算例及比较分析
4.1 数值算例
LSGDM的特点是由20人以上的专家组成的专家组，其目的是让专家达成共识，从多个备选方案中选择最佳方案，因此，作者采用图3的社交网络，选择25位专家、4个备选方案和3个属性来验证所提方法的有效性。
LSGDM is characterized by an expert group composed of more than 20 experts. The purpose is to allow experts to reach a consensus and select the best solution from multiple alternatives. Therefore, the author adopts the social network in Figure 3 to select 25 experts, 4 alternatives and 3 properties are used to verify the effectiveness of the proposed method.
接着作者按照前文所描述的关键步骤进行计算，考虑到供应商的选择问题，服装公司需要选择适合自己公司生产的材料。有四种材质可供选择，分别是:人造丝植绒、PVC植绒、针织布植绒、牛仔裤植绒。从以下三个属性来评价:美学，便利性和舒适度。
Then the author calculates according to the key steps described above. Considering the choice of suppliers, clothing companies need to choose materials suitable for their own production. There are four materials to choose from, namely: rayon flocking, PVC flocking, knitted fabric flocking, and jeans flocking. It is evaluated from the following three attributes: aesthetics, convenience and comfort.
随着分为四轮情况的计算，可以得到最终的计算结果，而每个迭代中共识级别的变化如表5所示，最后作者获得一个群体决策矩阵，和各方案的总体评价值，可选方案的衍生排序为x2>x4>x3>x1，因此，最适合的材料是x2。
With the calculation divided into four rounds, the final calculation result can be obtained, and the change of consensus level in each iteration is shown in Table 5. Finally, the author obtains a group decision matrix and the overall evaluation value of each scheme, optional The derivation order of the scheme is x2>x4>x3>x1, therefore, the most suitable material is x2.
4.2 比较分析
作者在本节通过对比分析，说明本文提出的方法的优点以及与其他方法的区别。使用4.1节的数值例子，将所提出的方法与TOPSIS(基于理想解相似度的顺序偏好技术)方法进行比较；当专家达成共识时，可以得到群体决策矩阵，然后可以计算出理想解的相似度，最后获得备用方案的排序：x2>x3>x4>x1。
In this section, the author explains the advantages of the method proposed in this paper and the difference from other methods through comparative analysis. Using the numerical example in Section 4.1, the proposed method is compared with the TOPSIS (Order Preference Technique Based on Ideal Solution Similarity) method; when the experts reach a consensus, the group decision matrix can be obtained, and then the similarity of the ideal solution can be calculated , and finally obtain the ordering of the alternatives: x2>x3>x4>x1.
接着作者在比较两种方法的备选方案排名和迭代次数后提出了2点意见：（1）虽然方案的排序与所提出的方法不同，但最合适的方案是x2。（2）当共识阈值θ设为0.75时，提出的方法需要2次迭代才能达成共识。但当共识阈值θ设置为0.8时，所提出的方法所需迭代次数为4次，TOPSIS方法为11次。此外，作者所提出的方法的共识速度比TOPSIS方法更快。
Then the author puts forward 2 opinions after comparing the ranking of alternatives and the number of iterations of the two methods: (1) Although the ranking of the schemes is different from the proposed method, the most suitable scheme is x2. (2) When the consensus threshold θ is set to 0.75, the proposed method requires 2 iterations to reach consensus. But when the consensus threshold θ is set to 0.8, the number of iterations required for the proposed method is 4, and 11 for the TOPSIS method. Furthermore, the consensus speed of the proposed method is faster than that of the TOPSIS method.
5. 仿真分析与讨论
5.1. 仿真分析
作者采用的CPR方法，将有界置信度和社交网络集成到反馈机制中，并给出仿真分析来验证所提出的方法，首先是构造算法，其次是模拟分析，最后是结果分析。仿真分析的目的是显示在专家的不同置信值和有界置信值下的每次迭代中的群体共识水平(见算法3)
The authors adopt a CPR approach that integrates bounded confidence and social networks into the feedback mechanism, and give simulation analysis to validate the proposed method, first by constructing the algorithm, second by simulation analysis, and finally by result analysis. The purpose of the simulation analysis is to show the level of community consensus in each iteration at different and bounded confidence values of the experts (see Algorithm 3)
5.2. 讨论
作者在此节讨论了本文中所提出的方法与现有研究的一些方法的不同之处，并将以往的文献提出的方法分为三类：T-LSGDM（传统LSGDM），SN-LSGDM（基于社交网络），BC and SN-LSGDM（基于有限置信度和社交网络）；表8显示了相关比较分析。
In this section, the author discusses the differences between the method proposed in this paper and some methods in existing research, and divides the methods proposed in the previous literature into three categories: T-LSGDM (traditional LSGDM), SN-LSGDM (based on social network), BC and SN-LSGDM (based on limited confidence and social network); Table 8 shows the relevant comparative analysis.
作者通过对比分析得出：（1）现有的T-LSGDM研究基于专家提供的偏好信息在大群体中研究CRP，这些研究假设专家彼此独立，没有考虑专家意见之间的相似性。这与本文的研究不同。相反，本文的研究不仅考虑了专家之间的相互信任关系，而且考虑了专家之间意见的相似性。（2）SN-LSGDM基于社交网络在大群体中研究CRP，在这些基于社交网络的反馈机制中，专家修改意见时应参考群体意见，而与专家在社交网络中的信任关系无关。他们的研究侧重于通过调整领导者的意见或在意见领袖之间增加连接边来达成共识，他们提出的方法与本文中提出的不同。（3）现有反馈机制假设反馈机制提供的建议必须为专家所接受。然而，如果建议与专家的意见相去甚远，专家可能会倾向于拒绝建议，不调整自己的意见，为了解决这一问题，作者在设计CRP中的反馈机制时考虑了有界置信模型。
Through comparative analysis, the authors concluded: (1) Existing T-LSGDM studies studied CRP in large groups based on preference information provided by experts, these studies assumed that experts were independent of each other, and did not consider the similarity between expert opinions. This is different from this study. On the contrary, the research in this paper considers not only the mutual trust relationship between experts, but also the similarity of opinions between experts. (2) SN-LSGDM studies CRP in large groups based on social networks. In these feedback mechanisms based on social networks, experts should refer to group opinions when revising their opinions, regardless of the trust relationship of experts in social networks. Their research focuses on reaching consensus by adjusting leaders' opinions or adding connecting edges between opinion leaders, and their proposed method differs from that presented in this paper. (3) The existing feedback mechanism assumes that the suggestions provided by the feedback mechanism must be accepted by experts. However, experts may be inclined to reject a recommendation and not adjust their own opinion if the recommendation is far from the opinion of the expert. To address this, the authors consider a bounded confidence model when designing the feedback mechanism in CRP.
总的来说，对于基于有界置信和社会网络的LSGDM问题，本文提出了一种新的反馈机制，该机制不仅考虑了专家之间的信任关系，而且考虑了专家的有界置信水平，这种方法为实际LS-GDM问题的求解提供了一种新的思路。
In general, for the LSGDM problem based on bounded confidence and social networks, this paper proposes a new feedback mechanism that considers not only the trust relationship between experts, but also the bounded confidence level of the experts, which This method provides a new idea for solving practical LS-GDM problems.
今天的分享就到这里了。
如果您对今天的文章有独特的想法，
欢迎给我们留言，
让我们相约明天，
祝您今天过得开心快乐！
That's it for today's sharing.
If you have a unique idea about today’s article,
Welcome to leave us a message,
Let us meet tomorrow,
I wish you a happy day today!
参考资料：谷歌翻译
参考文献：Li Yanhong et al. Consensus reaching process in large-scale group decision making based on bounded confidence and social network[J]. European Journal of Operational Research, 2022, 303(2) : 790-802.
本文由LearningYard学苑原创，如有侵权请在后台留言！
文案 |Yuan
排版 |Yuan
审核 |Qian
,
