Multi-input fuzzy-guided adaptive sine cosine algorithm for high-dimensional and constrained engineering optimization

Abstract The complexity, nonlinearity, and high-dimensional nature of modern engineering optimization problems present significant challenges for conventional metaheuristic algorithms. This study proposes a Fuzzy Tuned Sine Cosine Algorithm (FTSCA) that enhances the classical Sine Cosine Algorithm (SCA) using a Multi-Input Multi-Output Fuzzy Logic Controller (MIMO-FLC) for adaptive parameter regulation. The proposed fuzzy controller dynamically adjusts the inertia weight and tendency factor using three optimization indicators: normalized objective function, normalized distance from the best solution, and search tendency. Gaussian membership functions are employed for input fuzzification, triangular membership functions are used for the outputs, and 21 fuzzy inference rules govern the adaptive control process. The fuzzy outputs are incorporated into the SCA position update mechanism to dynamically regulate the search behaviour and improve the balance between exploration and exploitation. The proposed method is evaluated on 28 benchmark functions, the CEC2022 benchmark suite, scalability tests with dimensions 10, 50, and 100, the wind turbine blade design problem, and 16 constrained engineering optimization problems. Experimental results demonstrate that FTSCA consistently outperforms the conventional SCA and several improved SCA variants. It achieves the best average rank of 1.00 compared with 1.7857 for classical SCA across benchmark functions and estimates the global optimum on multiple test problems with very low standard deviation. Statistical validation using Wilcoxon, Friedman, and ANOVA tests confirms the significance and robustness of the obtained results. These findings indicate that FTSCA provides a robust, scalable, and efficient optimization framework for solving challenging high-dimensional engineering problems.