Hare escape optimization algorithm with applications in engineering and deep learning

The Hare Escape Optimization (HEO) algorithm is a novel metaheuristic inspired by the evasive movement strategies of hares when pursued by predators. Unlike conventional nature-inspired algorithms, HEO integrates Levy flight dynamics and adaptive directional shifts to enhance the balance between exploration and exploitation, improving its ability to escape local optima and converge efficiently. To validate its effectiveness, HEO was tested against 29 state-of-the-art metaheuristics on 43 benchmark functions from the CEC 2015 and CEC 2020 testbeds, demonstrating superior performance in both unimodal and multimodal landscapes. Beyond benchmark validation, HEO was applied to four complex constrained engineering design problems; spring, welded beam, pressure vessel, and truss optimization where it outperformed leading optimization methods in solution feasibility and computational efficiency. Additionally, HEO was employed to optimize hyperparameters in convolutional neural networks (CNNs) for image classification tasks, significantly enhancing model accuracy and convergence speed. The results indicate that HEO is a robust, adaptable optimization tool with promising applications in both engineering and deep learning. Its unique search mechanism provides a new perspective in metaheuristic optimization, opening pathways for further advancements in intelligent optimization techniques.