A novel marine predator whale optimization algorithm for global numerical optimization
Ya Su, Yi Liu
Abstract
Inspired by the horizontal pectoral herding tactics of humpback whales, this article presents a novel marine predator whale optimization algorithm (MPWOA), which incorporates predator–prey behaviours from the whale optimization algorithm (WOA) and marine predators algorithm (MPA). To mimic unique foraging behaviours, MPWOA uses three distinct motion models: Sinusoidal, Brownian and Lévy motions. Brownian and Lévy motions enhance the exploration and transition phases, while Sinusoidal motion ensures smooth transitions from exploration to exploitation, and assists in updating individual positions during the exploitation phase. The algorithm employs a ranking-based updating strategy, chaotic map sequences and adaptive inertia weights to improve performance further by avoiding local optima and accelerating convergence. Comprehensive evaluations on 39 CEC 2017/2019 benchmark functions and four real-world engineering optimization problems show that MPWOA exhibits competitive performance in providing faster convergence rates and higher-quality solutions over 12 state-of-the-art optimization algorithms and eight WOA variants, demonstrating effective exploration–exploitation balance.