An Anisotropic Tensor-Coupled Longhorn Beetle Optimization Algorithm with Multi-Field Adaptive Sensing Mechanism

To address the multimodality, dimensional coupling, and local extremum traps inherent in complex high-dimensional nonlinear optimization problems, an optimization algorithm based on the antennal sensing behavior and swarm collaboration mechanism of longhorn beetles is proposed. The algorithm uses an anisotropic antennal tensor as its core structure, modulates the search geometry using population statistical information, introduces a local curvature adaptive mechanism to regulate antennal extension and contraction, constructs a pheromone diffusion density field to achieve swarm collaboration, and implements adaptive state switching through an energy dissipation model. Furthermore, the algorithm is uniformly expressed as a stochastic dynamical system with multiple potential coupling. Theoretical analysis shows that this method can achieve a dynamic equilibrium between exploration and exploitation and exhibits asymptotic stability in a continuous time frame. The algorithm possesses strong structured modeling characteristics and good scalability.