Research on Node Robustness Based on Improved FW Algorithm

This paper proposes a two-layer robustness assessment model for urban road traffic systems. The upper layer establishes a robustness numerical model, while the lower layer builds a traffic flow distribution model for the road network, aiming to maximize robustness. The improved Frank-Wolfe algorithm based on node betweenness is used for the lower-layer traffic volume distribution, thereby optimizing the robustness of the road network. The change in the robustness value is observed to identify critical nodes in the network. The experiment selects the main roads within Beijing's Second Ring Road for analysis, revealing that the Frank-Wolfe algorithm, established with robustness maximization as the objective function, achieves higher distribution efficiency. The number of failed segments can be reduced by 30.7%• For the Second Ring Road network in Beijing, nodes such as Dongsi—located in the southern part of Beijing—are identified as critical nodes that have the greatest impact on the network's robustness, aligning with actual conditions. In summary, the robustness numerical model and the improved FW algorithm in this paper can better characterize the state of the road network under disturbances. As a metric in traffic volume distribution, it can more effectively identify critical nodes within the network.