[1]GAO Yanli,XIONG Zhihao,CHEN Shiming.Research on dynamic recovery strategies for edge-coupled interdependent networks[J].CAAI Transactions on Intelligent Systems,2024,19(1):238-248.[doi:10.11992/tis.202305019]
Copy
CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume:
19
Number of periods:
2024 1
Page number:
238-248
Column:
吴文俊人工智能科学技术奖论坛
Public date:
2024-01-05
- Title:
-
Research on dynamic recovery strategies for edge-coupled interdependent networks
- Author(s):
-
GAO Yanli; XIONG Zhihao; CHEN Shiming
-
College of Electrical and Automation Engineering, East China Jiaotong University, Nanchang 330013, China
-
- Keywords:
-
edge-coupled interdependent networks; cascading failure; recovery model; optimizing strategy; random failure; deliberate attack; Erd?s-Rényi random network; scale-free network
- CLC:
-
TP393
- DOI:
-
10.11992/tis.202305019
- Abstract:
-
To cope with the extensive structural damage or even collapse of the networks caused by a minor fraction of edge failures in edge-coupled interdependent networks, this paper proposes a cascading failure repair model for edge-coupled interdependent networks based on mutual boundary edges, dynamically integrating the repair process with the cascading failure process. And according to the features of edge-coupled network, a selective repair strategy based on compound excessive degree (SRCED) and an improved SRCED (ISRCED) are introduced. Simulation studies on Erd?s-Rényi random network (ER-ER) and scale-free network (SF-SF) edge-coupled interdependent networks are conducted under random failure and deliberate attacks. Comparative analysis with the randomly repair strategy (RR) and the selective repair strategy based on edge betweenness (SREB) is performed to find the optimal repair strategy. The research reveals that the optimal repair strategy in different structured edge-coupled interdependent networks changes depending on the initial edge retention ratio and the type of failure. Moreover, the repair strategy, capable of restoring the network to its initial state at a higher initial attack ratio of edges, does not necessarily have the least number of iterations.