Optimizing Network Toughness in Edge-Coupled Networks: BESCN Load Distribution Approach

In this paper, we explore the structural toughness of edge-coupled networks in communication systems, focusing on the impact of various attack strategies. We propose a general load distribution model for edge-coupled networks and analyze the impact of different attack patterns from the perspective of network load distribution. By considering three edge measurements-Jaccard coefficient, ED index, and EB coefficient-we optimize the network load distribution. First, we use Boundary-detection-based evolutionary stochastic configuration networks to generate an initial population that satisfies the supervisory mechanism by detecting potential interval boundaries. Subsequently, a Q-learning based selection strategy is used to adaptively select appropriate parameter settings for the evolutionary operator based on the interval boundaries where the population is located. Finally, experiments on a benchmark dataset demonstrate the tightness and accuracy of the proposed algorithm in constructing the model. Simulation results show that the proposed model enhances the network's toughness and reduces the impact of attacks compared to the traditional model. This study provides insights into the robustness of edge-coupled networks and offers strategies to improve network resilience.