IMPROVED SHARK SMELL OPTIMIZATION WITH ADAPTIVE PARAMETER SCHEDULING FOR FUZZY CONTROLLER TUNING

This paper proposes an Improved Shark Smell Optimization (ISSO) algorithm with adaptive parameter scheduling for optimal tuning of fuzzy controller parameters in inverted pendulum stabilization. Inspired by parameter control principles from particle swarm optimization, the algorithm introduces three linearly scheduled parameters: decreasing gradient weight, increasing inertia weight, and decreasing velocity limit. Applied to optimize a Fuzzy Hierarchical Swing-up and Sliding Controller (FHSSC), ISSO demonstrates statistically significant improvements over standard SSO and competitive performance with state-of-the-art metaheuristics. Experimental results show a 15.23% reduction in pendulum angle error (p<0.01), 26.72% reduction in cart position error (p<0.001), and 21.37% faster settling time compared to standard SSO. Systematic robustness analysis across 100 perturbed initial conditions confirms the controller’s stability where other methods fail. While the adaptation mechanism is conceptually adapted from PSO, its integration into SSO and validation on a complex control problem constitutes a meaningful contribution. The work provides both algorithmic refinements to SSO and practical validation for complex control system optimization.