运筹与管理 ›› 2024, Vol. 33 ›› Issue (6): 1-6.DOI: 10.12005/orms.2024.0173

• 理论分析与方法探讨 •    下一篇

装备保障网络级联失效及韧性分析

兑红炎1, 许哲1, 白光晗2, 陶俊勇2   

  1. 1.郑州大学 管理学院,河南 郑州 450001;
    2.国防科技大学 智能科学学院,湖南 长沙 410073
  • 收稿日期:2022-03-21 出版日期:2024-06-25 发布日期:2024-08-14
  • 通讯作者: 陶俊勇(1969-),男,内蒙古赤峰人,教授,博士,研究方向:复杂系统故障诊断和可靠性。
  • 作者简介:兑红炎(1982-),男,河南新郑人,教授,博士,研究方向:系统可靠性和重要度;许哲(1997-),男,河南商丘人,硕士,研究方向:系统可靠性;白光晗(1986-),男,贵州贵阳人,副研究员,博士,研究方向:复杂系统可靠性、抗毁性、稳健性分析。
  • 基金资助:
    国家自然科学基金资助项目(72071182)

Cascading Failure and Resilience Analysis of Equipment Support Network

DUI Hongyan1, XU Zhe1, BAI Guanghan2, TAO Junyong2   

  1. 1. School of Management, Zhengzhou University, Zhengzhou 450001, China;
    2. Intelligence Science College, National University of Defense Technology, Changsha 410073, China
  • Received:2022-03-21 Online:2024-06-25 Published:2024-08-14

摘要: 针对作战系统网络单元的异质性以及各个单元之间的交互联系,本文建立了包含作战和保障单元的三层装备保障多层耦合网络,将作战系统体系结构分为打击敌方目标的作战子系统、对我方毁伤装备和机构进行维修的保障子系统、对我方装备和机构进行物资供应的保障子系统。同时给出了基于网络拓扑结构和节点间信息传递类型的重要度评估方法。其次,基于负载-容量模型研究网络的级联失效传递过程,通过节点重要度进行蓄意攻击,判断负载再分配后的网络是否失效。然后,为了增强多层耦合网络韧性,基于剩余杀伤链数量,提出剩余可分配任务量韧性策略,缓解级联失效对整个网络的影响。最后,通过对不同攻击方式下多层耦合网络韧性进行仿真,验证了基于重要度进行攻击的效果明显优于随机攻击;对实施与未实施剩余可分配任务量韧性策略时的多层耦合网络韧性进行仿真,结果表明该策略能大幅提升网络韧性,减少级联失效的影响,增加网络在失效过程中的平均性能。

关键词: 可靠性, 重要度, 级联失效, 韧性

Abstract: Aiming at the heterogeneity of combat system network units and the interaction connection between each unit, this paper establishes a three-layer equipment safeguard multilayer coupling network containing combat and safeguard units, divides the network into kill, maintenance and storage and supply layers, and at the same time gives an importance assessment method based on the topology of the network and the type of information transfer between nodes. Secondly, the cascade failure transfer process of the network is studied based on the load-capacity model, and a cascade failure model of the equipment safeguard network is established. Finally, in order to enhance the multilayer coupled network toughness, the remaining assignable task volume toughness strategy is proposed. By simulating the multilayer coupled network toughness under different attack methods, it is verified that the effect of the attack based on the importance degree is significantly better than that of the random attack; the multilayer coupled network toughness with and without the implementation of the residual assignable task volume toughness strategy is simulated, and the results show that the strategy can significantly enhance the network toughness, reduce the impact of cascade failure, and increase the average performance of the network in the process of failure.
In this paper, the combat system architecture is divided into the combat subsystem for striking enemy targets, the support subsystem for repairing our destroyed equipment and institutions, and the support subsystem for supplying our equipment and institutions with materials. In order to elucidate the coupling relationship of different subsystems that are interconnected under the driving of combat mission tasks, the combat subsystem is set up as a kill layer (A), the maintenance subsystem as a repair layer (R), and the storage and supply subsystem as a storage and supply layer (S), which constitute the multiple coupling network of equipment safeguard through the command relationship and support synergies among each other (ψ).
In this paper, we analyze the cascade failure of the equipment assurance network based on the load-capacity model, and the specific failure process is as follows: 1)Assuming that the enemy carries out a deliberate attack based on the importance of our node (IXij), and the node fails after the attack. 2)Determining the initial load of the network node(RXij(0)). 3)Redistributing the load of the failing node. 4) Updating the node load, and determining whether the node fails after the load is redistributed. 5) Determining whether the network is stable.
In this paper, we propose the remaining assignable task amount toughness strategy, which is to make the neighboring node (original node) of the node reconnect to the same type of node of the failed node with the largest priority recovery factor (H) in the network with a certain probability after intentionally attacking a node; this strategy is implemented starting from the end of the cascade failures within the attacked network layer, and it aims at mitigating the impact of the cascade failures on the whole network. The ratio of the number of remaining kill chains (PNA*) is used as the network’s toughness indicator; the number of kill chains indicates the degree of redundancy in destroying the links to the target nodes; the higher the number of kill chains and the higher the redundancy, the lower the degree to which the network is affected and the higher the network toughness.
The simulation analysis is based on complex network visualization software Gephi generated by 90 nodes composed of equipment security multi-layer coupling network, and the simulation takes different information importance weight α=1/2,β=1/3,γ=1/6; node internal and external load factor σ=1,τ=1.2; node topology information load factor δ=0.3; load tolerance factor ω=0.2. Multi-layer coupled network toughness during attack is based on information importance and compared with random attack scenario. The comparison concludes that the residual assignment task resilience strategy proposed for the equipment support multi-layer coupled network model can greatly reduce the failure of the multi-layer coupled network after the attack, improve the resilience of the network, and increase the average performance of the network in the failure process.

Key words: reliability, importance measure, cascading failure, resilience

中图分类号: