Geomagnetic navigation metaheuristic algorithm for engineering optimization

This paper proposes a novel meta-heuristic optimization algorithm - Geomagnetic Navigation Algorithm (GNA), which is inspired by the geomagnetic navigation mechanism of migratory birds. The algorithm constructs three core mechanisms: (1) Geomagnetic gradient dominance and multi-source cognitive modulation mechanism, which maps the direction of the magnetic inclination gradient to the global optimal solution and integrates information from three channels: elite memory, group social cognition, and magnetic pole reinforcement attraction; (2) Adaptive cognitive landmark chain correction mechanism, which dynamically selects the top three individuals of the population as cognitive landmark chains, and their guiding weights are randomly generated each generation to simulate the reliability assessment of landmarks, achieving multi-scale references from near to medium to long distances; (3) Triple heavy-tailed distribution bionic perturbation mechanism, which switches between Cauchy/Gaussian/pseudo-Levy distributions with equal probability to simulate three types of environmental disturbances: geomagnetic storm impact, magnetic field micro-fluctuation, and terrain exploration jumps. In the comparison with six algorithms such as WOA, TOC, and SCA on the CEC-2017 test set of 10/30/50 dimensions and the verification of engineering optimization problems such as pressure vessel design, GNA significantly outperforms the comparison algorithms. Experiments confirm that GNA has significant advantages in the balance of exploration and exploitation, resistance to local optimality, dimension adaptability, and handling of engineering constraints, providing an efficient solution for complex optimization problems.