Resilience Evaluation and Optimization of UAV Swarm Network under Low Probability Risk

This paper addresses the process description issue of topology changes in Unmanned Aerial Vehicle (UAV) Mobile Ad Hoc Networks (UMANETs) caused by low probability risk and proposes an approach to evaluating network resilience. Assessing network resilience differs from evaluating network robustness and reliability against common high-probability risks. Instead, network resilience refers to the network’s capacity to withstand damage and recover from low-probability and severe risks. The concept of network resilience using process modeling, requirement analysis, and structural optimization is presented. The paper provides criteria for evaluating network resilience in four aspects, including network reconstruction capability, communication efficiency, chain break rate, and the degree of network loading. Dynamic Stackelberg game theory is utilized to simulate the conflict between network disruptions and a real-time planning system, analyzing the dynamic evolution of topology. To dynamically plan a more resilient network topology in real-time, the proposed approach introduces an enhanced Ant Colony Algorithm that leverages a Bio-inspired Neural Network (BNNAC). Finally, simulation results demonstrate the effectiveness and superiority of the proposed scheme.