PID controller parameter optimization for ship fin stabilizers based on an improved secretary bird optimization algorithm

Abstract

Fin stabilizers governed by proportional-integral-derivative (PID) controllers dominate active ship roll reduction, yet their performance depends critically on gain tuning under nonlinear hydrodynamics and time-varying sea states. This paper proposes an Improved Secretary Bird Optimization Algorithm (ISBOA) combining a Logistic-Tent chaotic map with opposition-based learning, a DE/rand/1/bin operator in the predation phase, and a nonlinear adaptive control factor with energy-driven stagnation escape. Across 29 CEC2017 and 12 CEC2022 functions ISBOA ranks first overall; of 246 pairwise comparisons 199 are significant and 173 also reach medium or large favourable effect size, the exception being CCO. Ablation confirms that each strategy contributes. Applied to PID tuning in a Matlab/Simulink closed loop driven by a Pierson-Moskowitz spectrum, minimizing a roll-angle ITAE index, ISBOA reduces RMS and peak roll by 95.3% and 94.4% against 85.6% and 82.7% for a conventionally tuned fixed PID. With the gains fixed across nine off-design conditions the reduction stays within 94.2–95.9% in eight, but the loop diverges in the ninth; a coefficient sweep shows the tuning is stable only to 0.90 of nominal whereas lower-gain settings tolerate 0.75, an explicit performance/stability-margin trade-off. Test bench measurements show peak pipeline vibration falling 25.2 dB.