Spotted Shark Optimization Algorithm

To address the common problems in high-dimensional complex optimization problems, such as premature convergence, unbalanced search direction, and insufficient global exploration capability, this paper proposes a spotted shark optimization algorithm based on an information speckle tensor field and a self-organizing critical energy breaking mechanism. The algorithm constructs a unified coupled model from three levels: information field, energy field, and topological evolution. First, an information speckle tensor is constructed to achieve anisotropic search control. Second, a critical energy breaking mechanism based on population energy variance is introduced to achieve self-organizing transformation in the search phase. Then, a direction correction model driven by local curvature is established to improve convergence efficiency on complex surfaces. Furthermore, a topological speckle density self-evolution mechanism is designed to dynamically couple search intensity with local fitness structure. Finally, a search space memory folding operator is introduced to avoid repeated visits to inferior regions. Theoretically, the algorithm behaves as a stochastic dynamical system with tensor preconditions, exhibiting weak Newtonian approximation characteristics and a phase transition driving mechanism. This paper presents the complete mathematical model and update equations, and analyzes its dynamic properties and complexity.