A modified black-winged kite optimizer based on chaotic maps for global optimization of real-world applications

Optimization algorithms play a critical role in solving complex engineering and medical imaging optimization problems. However, existing metaheuristic techniques often suffer from premature convergence, inefficient exploration, and imbalance between exploration and exploitation. To address these limitations, this paper proposes the Modified Black-Winged Kite Optimizer (M-BWKO), an enhanced version of the standard BWKO algorithm. M-BWKO incorporates six key improvements: a top-k elite leader strategy, adaptive chaos weighting, diversity-aware chaos reactivation, chaotic index-based selection, adaptive Cauchy mutation, and a hybrid migration rule combining chaotic perturbations, Cauchy mutation, and directional updates. The selected M-BWKO variant, Tent-BWKO (TT-BWKO), is evaluated on the CEC-2022 benchmark suite, achieving up to 22.04 % improvement over BWKO and 99.99 % over other state-of-the-art optimizers, with average gains of 6.30 % and 22.13 %, respectively. These results are statistically validated using the Wilcoxon rank-sum test (p < 0.05), confirming the robustness of the approach. TT-BWKO is further tested on real-world engineering design problems—including Welded Beam, Tension/Compression Spring, and Pressure Vessel—resulting in notable reductions in material cost. It also performs effectively on large-scale Traveling Salesman Problem instances (100, 150, 200 cities), demonstrating strong route optimization and stability. In medical image segmentation, TT-BWKO yields superior PSNR, SSIM, and FSIM scores, confirming its versatility and effectiveness across diverse domains.