| 引用本文: | 刘凤增,肖兵,刘俊杰,等.级联失效下非对称依赖网络鲁棒性研究.[J].国防科技大学学报,2021,43(1):49-56.[点击复制] |
| LIU Fengzeng,XIAO Bing,LIU Junjie,et al.Robustness of asymmetric dependent network under cascading failure[J].Journal of National University of Defense Technology,2021,43(1):49-56[点击复制] |
|
| |
|
|
| 本文已被：浏览 7482次 下载 5229次 |
| 级联失效下非对称依赖网络鲁棒性研究 |
| 刘凤增1,2,肖兵1,刘俊杰2,王昊冉3 |
|
(1. 空军预警学院 预警情报系, 湖北 武汉 430019;2. 国防科技大学 信息通信学院, 湖北 武汉 430010;3. 国防科技大学 系统工程学院, 湖南 长沙 410073)
|
| 摘要: |
| 研究级联失效下网络化信息物理系统的鲁棒性具有重要的现实意义。针对具有单向依赖关系的信息物理系统构建双层非对称依赖网络模型,综合依赖失效和过载失效设计级联失效模型,提出一种更贴近实际的非对称攻击方式,并给出一种资源限制下的节点容量分配方式。针对无标度子网构成的非对称依赖网络进行了仿真实验,结果表明:网络鲁棒性与子网络平均度、度指数、节点容量等呈正相关关系；同配依赖网络以及按度值分配节点容量的网络具有更高的鲁棒性；非对称依赖网络面对蓄意攻击时具有脆弱性；以同等力度单独攻击被依赖子网时,网络受到的损伤更大。构建的模型及发现的规律对于研究网络鲁棒性、优化网络设计具有一定的参考价值。 |
| 关键词: 非对称依赖网络 鲁棒性 级联失效 过载失效 |
| DOI：10.11887/j.cn.202101007 |
| 投稿日期：2019-07-31 |
| 基金项目:国家自然科学基金资助项目(61502522)；国家部委基金资助项目(JZX7Y20190253036101) |
|
| Robustness of asymmetric dependent network under cascading failure |
| LIU Fengzeng1,2, XIAO Bing1, LIU Junjie2, WANG Haoran3 |
|
(1. Department of Early Warning Intelligence, Air Force Early Warning Academy, Wuhan 430019, China;2. College of Information and Communication, National University of Defense Technology, Wuhan 430010, China;3. College of Systems Engineering, National University of Defense Technology, Changsha 410073, China)
|
| Abstract: |
| It is of great practical significance to study the robustness of networked CPS(cyber physical systems) under cascading failure. A two-tier asymmetric dependency network model was constructed for CPS with one-way dependency and a cascading failure model for asymmetric dependency network which combines the dependency failure with the overload failure was designed. A more practical asymmetric attack method was proposed, and a node capacity allocation method under resource constraints was given. Simulation experiments were carried out on asymmetric dependent networks consisting of scale-free subnets. It is found that the network robustness is positively correlated with the sub-network average, degree exponent, node capacity, etc.; assortative dependency networks and networks with node capacity allocation by degree are more robust; asymmetric dependent networks are vulnerable to attention attacks; when attacking dependent subnets with equal strength, the network is more vulnerable. The constructed model and the law of experimental discovery have certain reference value for studying network robustness and optimizing network design. |
| Keywords: asymmetric dependent network robustness cascading failure overload failure |
|
|
|
|
|
