基于复杂网络的中国航空网络研究
发布时间：2018-07-21 20:26
【摘要】：当今时代是信息时代。复杂网络学科因为信息化的爆炸性发展成为了近年来新兴的研究领域。随着社会的不断发展,使用网络拓扑结构不但能够解决现实社会中的实际问题,同时也可以使问题更加系统化、直观化。通过加深对复杂网络的研究,可以使网络更好的服务人类,也使人类对网络的认识更加深刻,推动复杂网络的进一步发展。第一,运用复杂网络知识分析了中国航空网络的基本结构。运用space L和space P两种建模方式构建ANL航空网络和ANP航空网络。分析结果证明了ANL航空网络和ANP航空网络具备复杂网络的基本特性;中国航空网络的吞吐量和航距与度数的关系都是幂律分布;中国航空网络中介数越大的机场,吞吐量和航距越大,但相比度值对这两个因素的影响,介数相关系数较小,而且并不是吞吐量越大,该节点在网络中发挥的作用就越大。第二,运用聚类算法对中国航空网络进行了分析。针对传统k-means算法随机选择初始中心所带来的问题,本文在去除孤立点的前提下选择距离乘积最远的点作为初始中心,这种改进明显提高了算法的准确率和稳定性,聚类质量也得到了显著提升。在用经典数据集验证改进算法的有效性后,将算法运用到航空网络中去,结果证明航空网络具有非均匀分布的无标度特性和中心化倾向。第三,分别删除单个节点和连续删除多个节点,运用网络效率、聚集系数、最大连通子图相对大小来分析中国航空网络的抗毁性。结果显示:航空网络中,重要机场在影响连通性方面发挥着至关重要的作用;当删除节点个数超过10个时,全局网络效率下降超过50%;当被攻击机场个数超过44个时,网络的全局效率下降99%,因此我们可以得知,中国航空网络抵御蓄意攻击的能力十分薄弱;比较按度攻击和按介数攻击后的指标发现,后者网络受到的破坏更大,说明在影响网络连通性这一方面介数比度值发挥的作用大。最后结合基础的测试指标提出一种新的抗毁性指标iA)(,结果显示新的测试指标更接近于现实,效果更好,能够更全面的考虑三个测试指标,不像传统算法具有片面性。第四,分析了基于社团结构的中国航空网络抗毁性。本文分别提出了基于社团结构的新攻击方式和基于社团结构的抗毁性测试指标,即考虑了网络的整体性能,也考虑到了网络的社团结构变化,并且用经典网络验证了攻击方式和测试指标的有效性。然后用不同方式攻击中国航空网络,我们可以得到:相比较传统攻击方式,基于社团的新攻击方式更容易而且更大程度的破坏了网络社团的结构;新的攻击方式对网络的稳定性也有一定的影响;中国航空网络的每个社团中都有介数较大的点,从一定程度可以证明存网络中存在关键机场,所以为了提高网络在受到自然灾害、人为攻击时的抗毁性,我们就要加强这些枢纽机场以及其备降机场的建设。本文的主要工作是运用复杂网络知识对构建的中国航空网络进行分析,为优化航空网络线路提供可参考的建议。
[Abstract]:The modern era is the information age. The complex network discipline has become a new research field in recent years because of the explosive development of information. With the continuous development of the society, the use of network topology can not only solve practical problems in the real society, but also make the problem more systematic and intuitive. By deepening the complex network The research can make the network better serve the human being, and make the human understanding of the network more profound and promote the further development of the complex network. First, the basic structure of China's aviation network is analyzed with complex network knowledge. The two modeling methods of space L and space P are used to construct the ANL aviation network and ANP aviation network. It is clear that the ANL aviation network and the ANP aero network have the basic characteristics of complex networks; the relationship between the throughput and the distance and degree of the Chinese aeronautical network is a power law distribution; the greater the number of airports in the China aviation network, the greater the throughput and the distance, but the relative coefficient of the two factors is smaller than the degree of comparison, and it does not. The greater the throughput, the greater the role of the node in the network. Second, using clustering algorithm to analyze the Chinese aeronautical network. Aiming at the problem caused by the random selection of the initial center of the traditional K-means algorithm, this paper selects the farthest point of the distance product as the initial center with the removal of the isolated point, and this improvement is obvious. The accuracy and stability of the algorithm are improved, and the clustering quality has been greatly improved. After the validity of the improved algorithm is verified by the classic dataset, the algorithm is applied to the aero network. The results show that the network has the nonuniform distribution of the scale-free characteristics and the centralization direction. Third, the individual nodes are deleted and the continuous deletion is deleted respectively. Multiple nodes, using network efficiency, aggregation coefficient, and the relative size of the Dalian pass graph to analyze the resistance of China aviation network. The results show that the important airport plays a vital role in influencing connectivity in the aviation network; when the number of deleted nodes is over 10, the efficiency of the global network drops by more than 50%; when attacked the airport When the number is more than 44, the global efficiency of the network is reduced by 99%. Therefore, we can learn that the capability of China aviation network to resist deliberate attack is very weak. Finally, a new anti destruction index iA is proposed based on the basic test index. The results show that the new test index is closer to the reality, the effect is better, and the three test indexes can be considered more comprehensively. Unlike the traditional algorithm, it is not one-sided. Fourth, the damage resistance of China aviation network based on community structure is analyzed. The new attack mode based on community structure and the destructive testing index based on community structure, that is, considering the overall performance of the network, and taking into account the changes in the community structure of the network, and using the classical network to verify the effectiveness of the attack mode and the test index. Then, we can use different formula to attack China's aviation network, we can get the phase. Compared with the traditional attack mode, the new attack mode based on the community is easier and more destructive to the structure of the network community; the new attack mode also has a certain influence on the stability of the network; there are a large number of points in each association of the Chinese aviation network, which can prove the existence of the key airport in the network to a certain extent. In order to improve the network's destruction resistance to natural disasters and human attacks, we should strengthen the construction of these hubs and its Alternate airport. The main work of this paper is to analyze the Chinese aviation network constructed with complex network knowledge and provide some suggestions for optimizing the network lines.
【学位授予单位】：青岛理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5;TP311.13
[Abstract]:The modern era is the information age. The complex network discipline has become a new research field in recent years because of the explosive development of information. With the continuous development of the society, the use of network topology can not only solve practical problems in the real society, but also make the problem more systematic and intuitive. By deepening the complex network The research can make the network better serve the human being, and make the human understanding of the network more profound and promote the further development of the complex network. First, the basic structure of China's aviation network is analyzed with complex network knowledge. The two modeling methods of space L and space P are used to construct the ANL aviation network and ANP aviation network. It is clear that the ANL aviation network and the ANP aero network have the basic characteristics of complex networks; the relationship between the throughput and the distance and degree of the Chinese aeronautical network is a power law distribution; the greater the number of airports in the China aviation network, the greater the throughput and the distance, but the relative coefficient of the two factors is smaller than the degree of comparison, and it does not. The greater the throughput, the greater the role of the node in the network. Second, using clustering algorithm to analyze the Chinese aeronautical network. Aiming at the problem caused by the random selection of the initial center of the traditional K-means algorithm, this paper selects the farthest point of the distance product as the initial center with the removal of the isolated point, and this improvement is obvious. The accuracy and stability of the algorithm are improved, and the clustering quality has been greatly improved. After the validity of the improved algorithm is verified by the classic dataset, the algorithm is applied to the aero network. The results show that the network has the nonuniform distribution of the scale-free characteristics and the centralization direction. Third, the individual nodes are deleted and the continuous deletion is deleted respectively. Multiple nodes, using network efficiency, aggregation coefficient, and the relative size of the Dalian pass graph to analyze the resistance of China aviation network. The results show that the important airport plays a vital role in influencing connectivity in the aviation network; when the number of deleted nodes is over 10, the efficiency of the global network drops by more than 50%; when attacked the airport When the number is more than 44, the global efficiency of the network is reduced by 99%. Therefore, we can learn that the capability of China aviation network to resist deliberate attack is very weak. Finally, a new anti destruction index iA is proposed based on the basic test index. The results show that the new test index is closer to the reality, the effect is better, and the three test indexes can be considered more comprehensively. Unlike the traditional algorithm, it is not one-sided. Fourth, the damage resistance of China aviation network based on community structure is analyzed. The new attack mode based on community structure and the destructive testing index based on community structure, that is, considering the overall performance of the network, and taking into account the changes in the community structure of the network, and using the classical network to verify the effectiveness of the attack mode and the test index. Then, we can use different formula to attack China's aviation network, we can get the phase. Compared with the traditional attack mode, the new attack mode based on the community is easier and more destructive to the structure of the network community; the new attack mode also has a certain influence on the stability of the network; there are a large number of points in each association of the Chinese aviation network, which can prove the existence of the key airport in the network to a certain extent. In order to improve the network's destruction resistance to natural disasters and human attacks, we should strengthen the construction of these hubs and its Alternate airport. The main work of this paper is to analyze the Chinese aviation network constructed with complex network knowledge and provide some suggestions for optimizing the network lines.
【学位授予单位】：青岛理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O157.5;TP311.13
【参考文献】
相关期刊论文 前10条
1 李武;赵娇燕;严太山;;基于平均差异度优选初始聚类中心的改进K-均值聚类算法[J];控制与决策;2017年04期
2 唐亮;;一种改进的DBSCAN算法[J];电脑知识与技术;2017年06期
3 曹卫华;乔平安;;改进K-Means算法的探讨与分析[J];电脑知识与技术;2017年06期
4 王菲菲;李秦;张梦佳;;k-means聚类算法的改进研究[J];甘肃科技纵横;2017年03期
5 赵文涛;赵好好;孟令军;;基于相关拓扑势的社团发现算法[J];计算机应用与软件;2017年01期
6 赵文涛;赵好好;孟令军;;基于节点内聚系数的局部社团发现算法[J];计算机应用与软件;2016年12期
7 伍恒;李文杰;蒋e，
本文编号：2136758
本文链接：https://www.wllwen.com/shoufeilunwen/xixikjs/2136758.html
