Resilient Topological Control for Dynamic Underwater Optical Wireless Networks

Underwater Wireless Optical Networks (UWONs) play a critical role in tasks such as ocean monitoring and resource exploration, which require high connectivity and reliability. However, water turbidity, ocean currents, and ambient light noise significantly affect communication stability, creating serious deployment challenges. To address this, we propose a network topology optimization method based on resilience evaluation. First, a resilience evaluation method is designed to quantify the network's ability to adapt to disturbances. Then, an improved predecessor-based evolutionary algorithm (Pred-EA) is used for resilience-guided topology optimization. To improve algorithmic efficiency and search quality, the prim algorithm is introduced to ensure chromosome feasibility, enhancing both the diversity of the initial population and computational efficiency. Experimental results show that our method achieves better recovery performance than three comparison methods in all scenarios. The average number of recovered edges improves by up to 18.30% over the second-best method. Under non-recoverable conditions, resilience improves by up to 20.52%. These results confirm the strong topological robustness and practical value of the proposed method in dynamic underwater environments.