Enhanced robustness of flow networks with dependency groups.

Dependency links represent the relationships between network nodes that have an interactive impact on cascading failures caused by load fluctuation in the network. However, existing research mainly focuses on load fluctuation's failure mechanisms without considering the dependency links of nodes and their cascading prevention mechanisms in reality. This study addresses the cascading prevention problem in networks when dependency links and connectivity links operate together. It proposes a hybrid cascading failure model based on the dependency relationships, load fluctuation and reinforced nodes. Furthermore, it provides four reinforced nodes' strategies that leverage static and local information characteristics of network nodes. These strategies help the network to perform its function and prevent cascading failures effectively. The study considers actual situations where overloaded nodes can still maintain their function. To measure the overload ability and the uncertainty of node failure, the authors used the overload coefficient parameter and the failure probability. Additionally, the impact of the dependency group's size on the network robustness is explored. Simulation results on BA and ER networks and two actual networks show that reinforced nodes' strategies provide significant support in keeping the network away from abrupt collapses. [ABSTRACT FROM AUTHOR]