Research On Robustness And Interdependency Models Of Critical Infrastructure

Critical Infrastructures(CIs)are the cornerstone of social prosperity and development.The safe and reliable operation of CIs is vital to the safety and stability of our society.However,CI is faced with many challenges,such as disasters,misoperations and cyber attacks.Meantime,the network convergence brings the risk of cascading failures.Robustness,which is defined as the survivability under failure or attacks,is one of the most important properties of CI.This thesis is dedicated to the evaluation and enhancement of the robustness of CI.The main content of this thesis includes:1.The interdependency model of CI: This thesis proposes a capability-based dependency model of CI that takes two node dependency properties into account.One is support-capability and the other is required-capability.The redundancy degree of CI is also defined,whose value is the ratio of total support capability to total required capability.Through the numerical simulations,we find that:(1)CI without redundant supportdependence links is extremely vulnerable,even the failure of one node could cause the collapse of whole CI;(2)increasing support-dependence links and redistributing the node's dependency properties can enhance the robustness of CI without changing its redundancy degree;(3)improving the redundancy degree could enhance network robustness without adding support-dependence links.These conclusions enlighten the design of CI: when CI's redundancy degree is fixed,we can add dependency links to enhance robustness;when CI's structure is settled,we can improve the redundancy degree to enhance robustness.2.The threat model of CI: CI is under threat from multiple kinds of attacks.Based on the objects and sequence of attacks,we divide the attacks into three kinds: simultaneous attack,sequential attack and composite attack.Through numerical simulations,we found that:(1)the sequential attack may bring more damages in single physical systems;(2)the coupling process of cyber system and physical system makes it possible that sequential attack brings more damages than simultaneous attack when the attackers only attack the cyber system;(3)with some target sets,composite attack leads to more failures than both simultaneous attack and sequential attack.The above results suggest that defenders should take all the three kinds of attacks into account when they select the critical nodes from CI.3.Link addition strategies for enhancing the robustness of CI network: This thesis investigates the link addition strategies considering both connectivity links and dependency links to enhance the robustness of CI.Enhancing robustness of CI by only adding connectivity links or dependency links has been researched extensively,however,few of them are focusing on adding both connectivity and dependency links to enhance the robustness.This thesis proposes a strategy to allocate the limited costs reasonably to add both connectivity and dependency links.The experiment results show that compared to link addition of the past,our link addition strategies could bring more robustness to CI.4.The self-healing mechanism based on swarm intelligence for CI network:This thesis proposes a self-healing mechanism of CI based on swarm intelligence,and the power grid is used for the case study.Load scheduling of thermostatical appliances is used to deal with the abrupt changes in a power grid.For the first time,we put forward the measurement of user-friendliness.Meantime,elastic T algorithm is proposed to stabilize the power grid.The experimental results show that our mechanism is user-friendly and can effectively adjust the state of the power grid so that the power grid can run safely and stably continuously.