Dynamic random walk-based sled dog optimization algorithm and artificial neural network for optimizing design engineering problems

Abstract

This research presents a modified version of the sled dog optimizer (SDO) to enhance optimization performance across various benchmark functions and real-world applications. The proposed modification introduces adaptive mechanisms to balance exploration and exploitation, thereby improving convergence speed and solution accuracy. Experimental results demonstrate that the modified SDO outperforms the standard SDO and other contemporary metaheuristic algorithms in terms of optimization efficiency and robustness. Comparative analysis of standard test functions and engineering design problems confirms the superiority of the proposed approach.