A Novel Starfish Optimization Algorithm for Secure STAR-RIS Communications

This paper develops an intelligent Enhanced Starfish Optimization (ESFO) algorithm for optimizing a secure wireless communication infrastructure. The Starfish Optimization (SFO) algorithm is inspired by starfish biology, using the integrated modeling of the arm-based exploration, preying, and regeneration behaviors of starfish. To further enhance the exploitation capability of the standard Starfish Optimization (SFO), the proposed Enhanced Starfish Optimization (ESFO) integrates a fitness-based interacting mechanism within the exploitation phase. This innovative modification improves local search accuracy, preserves population diversity, and mitigates premature convergence without introducing additional control parameters. Moreover, the proposed Enhanced Starfish Optimization (ESFO) is designed for secure wireless transmission, which is considered one of the main topics in next-generation wireless network infrastructure. The investigated network addresses the use of Simultaneously Transmitting and Reflecting RIS (STAR-RIS) in the security of the physical layer. This implemented STAR-RIS has a coupled phase shift to create reflected and transmission links, unlike traditional Reconfigurable Intelligent Surface (RIS). In this regard, we create a safe beamforming architecture that optimizes both Base Station (BS) precoding vectors and STAR-RIS transmission/reflection coefficients. In order to validate the efficiency of the proposed Enhanced Starfish Optimization (ESFO) algorithm, it is compared to several benchmark optimizers such as standard Starfish Optimization (SFO), Dhole Optimizer (DO), Neural Network Algorithm (NNA), Crocodile Ambush Optimization Algorithm (CAOA), and white shark Optimizer (WSO). These comparisons include several scenarios based on the transmitted power threshold which is varied in the range of 20 to 70 dBm with step of 5 dBm. The simulation results show that the proposed Enhanced Star Fish Optimization (ESFO) algorithm consistently outperforms existing benchmark approaches. This study supports future intelligent communication infrastructures in terms of secrecy and achievable rates over a range of transmit power levels. In particular, ESFO improves performance by up to 20-25% while converging 40-50% faster than traditional optimization algorithms, demonstrating its usefulness and resilience in STAR-RIS-assisted secure communication systems. The suggested ESFO-enabled architecture outperforms standard RIS-based systems in terms of secrecy capacity, according to numerical studies, and low-resolution STAR-RIS phase-shifters are sufficient to ensure robust secrecy performance.