Abstract
The sine cosine algorithm (SCA), a recently discovered population-based optimization technique, is used to resolve optimization issues. In this research, the study proposes employing the LWSCA (Locally Weighted Sine Cosine Algorithm) as a hybrid approach to enhance the performance of the original SCA (Sine Cosine Algorithm) and mitigate its limitations. These limitations encompass restricted resolution, slow convergence rates, and difficulties in achieving global optimization when dealing with complex, multi-dimensional spaces. The fundamental idea underlying LWSCA is to incorporate the SCA algorithm with the locally weighted (LW) technique and mutation diagram. The hybridization process has two stages: An algorithm is initially changed by altering the fundamental equations to ensure greater effectiveness and accuracy. The second point is that when the LW local approach is used to create a new dependent site, it increases the randomness during the search process. This, in turn, raises the population variance of the optimizer being proposed, ultimately enhancing the overall effectiveness of the global search. The putative method's hybrid architecture is anticipated to significantly increase the potential for exploration and exploitation. By evaluating SCA's performance against IEEE CEC 2017 functions and contrasting it with a variety of different metaheuristic techniques, the usefulness of SCA is investigated. According to the experimental data gathered, the LWSCA's convergence, exploration, and exploitation tendencies have all greatly improved. According to the results, the suggested LWSCA method is a good one that performs better than SCA and other rival algorithms in most functions.