Dynamic Adjustment of an Artificial Fish Swarm Algorithm Using Shadowed Type-2 Fuzzy Logic for Enhancing Benchmark Function Optimization

This paper presents a new improvement of the existing fuzzy artificial fish swarm algorithm (FAFSA). The main goal of this improvement is the incorporation of the shadowed type-2 fuzzy logic theory into the Artificial Fish Swarm Fuzzy Algorithm (SFAFSA), with the goal of dynamically adjusting the algorithm parameters as the iterations progress. The integration of shadowed type-2 fuzzy logic enables more adaptive and responsive behavior of the algorithm, leading to potentially improved performance and convergence. The main motivation behind the application of the shadowed type-2 fuzzy logic concept lies in its ability to handle uncertainties and variations more effectively. By dynamically adapting algorithm parameters across iterations, the FAFSA is better equipped to navigate complex and evolving optimization landscapes. This adaptability can potentially lead to more efficient and accurate solutions to various optimization problems. To evaluate the effectiveness of the proposed improvement, a series of experiments were performed. Reference mathematical functions of different dimensions were selected for optimization. The FAFSA was administered with integrated type 2 shaded fuzzy logic and the results were compared with those obtained using the traditional FAFSA and other optimization methods previously documented in the literature. The comparative analysis sheds light on the strengths and weaknesses of the different algorithms, offering information on their respective performances in different scenarios. Experimental results demonstrate promising improvements using the enhanced FAFSA. In particular, the algorithm's ability to dynamically adjust its parameters in response to the changing landscape provides advantages in terms of convergence speed and accuracy. These findings contribute to the growing body of research on optimization algorithms and highlight the potential of integrating advanced fuzzy logic concepts to improve algorithmic adaptability.