Exploring cascading failure in directed weighted small-world network: An adaptive SIRS-F mechanism

Abstract

Cascading failures in complex networks is a significant threat to system reliability and robustness, influencing various aspects of daily life. The small-world characteristics within complex networks contribute to increased complexity, making it challenging to predict the cascading failure due to the relationships and interaction among nodes. We assume that both node attacked and mutation can trigger cascading failure, and propose an adaptive SIRS-F mechanism based on load-capacity model and SIRS model to explore cascading failures in directed weighted small-world networks. In our mechanism, node loads continuously change due to a flow redistribution strategy, while parameters such as the infected rate and recovery rate of nodes adaptively vary. Through simulation, it was found that, compared to the SIRS model and the load capacity model, the cascading failure effects in directed weighted small-world networks under proposed mechanism align more closely with the ”avalanche” phenomenon. The network cascading failures conclude within $20<t<30$. This observation is more in line with real-world situations. Furthermore, we explored scenarios of cascading failures caused by node attacked and node mutations, and conducting sensitivity analysis to explore the influence of different parameters on the maximum collapse time of the network. The results indicate that cascading failures resulting from node attacked lead to a faster network collapse. Finally, case analysis results from power grid demonstrate the effectiveness of the proposed adaptive SIRS-F mechanism in ensuring network stability and optimizing network performance.