System recovery with memory after intentional attack: a simulation study
Network models are widely used to solve the practical system problems as the result of rapid development of network theory. Modern power system and communication system are both established on the basis of network theory. During the process of many natural network formation, scientists find the pref...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis-Master by Coursework |
| Language: | English |
| Published: |
Nanyang Technological University
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/159407 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Network models are widely used to solve the practical system problems as the result of rapid development of network theory. Modern power system and communication system are both established on the basis of network theory. During the process of many natural network formation, scientists find the preferential connection phenomenon. And this phenomenon will lead to that degree distributions of networks follow power-law distribution and these networks are known as scale-free networks. In scale-free networks, there are a small number of vertices with quite high degree and most vertices only have low degree. So the attack to these networks is often intended to destroy the key vertices in the networks. Because the connectivity and stability of this kind of networks mainly depend on these key vertices. In this dissertation, we study the recovery strategy of scale-free networks after massive intentional attack. For scale-free networks generation, we remove the correlation between vertices in the network and the degree distribution can be prescribed. We will evaluate the stability of networks with various degree distributions when facing attack. For recovery of networks, several different kinds of strategies will be studied. These strategies create new edges between the remaining vertices in the network and the connectivity and completeness of the networks can be measured with average distance and largest cluster size respectively. By comparing various repair methods, we hope improvement can be made to these methods and this dissertation can provide references for further research on networks recovery. |
|---|