Opposition-based starfish optimization algorithm for optimal design of digital FIR filters

This study proposes an enhanced variant of the Starfish Optimization Algorithm (SFOA), termed the Opposition-based Starfish Optimization Algorithm (O-SFOA), for the optimal design of digital FIR filters. Although SFOA is an effective meta-heuristic inspired by the exploration, hunting, and regeneration behaviors of starfish, it may suffer from premature convergence, stagnation, and entrapment in local minima. To alleviate these limitations, O-SFOA integrates three complementary mechanisms: Opposition-Based Learning (OBL) to strengthen global exploration, a dual regeneration strategy to improve escape from local optima, and a Gaussian-Based Restart (GBR) mechanism to preserve population diversity and enhance convergence robustness. The proposed O-SFOA is applied to the optimal design of 20-coefficient low-pass (LP), high-pass (HP), and band-pass (BP) FIR filters and compared with six meta-heuristic algorithms (COA, CSA, FA, GWO, MFO, and the original SFOA) over 30 independent runs. The experimental results show that O-SFOA is competitive across all filter types. For the LP filter, the maximum stopband attenuation reaches 159.738458 dB and the minimum passband ripple is 0.017703. For the HP filter, the maximum attenuation is 138.943226 dB with an average stopband ripple of 0.006992. For the BP filter, the maximum attenuation is 154.187850 dB and the average passband ripple is 0.040535. These results indicate that the proposed improvements significantly enhance the optimization performance and reliability of SFOA for digital FIR filter design.