26 October 2023 in IEEE Access
Optimization challenges are becoming more complex as the world advances. Since deterministic and heuristic approaches are no longer sufficient to deal with such complex problems, metaheuristics have recently emerged as a viable option to address optimization difficulties. Since Sand Cat Swarm Optimization (SCSO) is a famous meta-heuristic algorithm, SCSO has a weak ability to balance search between exploration and exploitation and slow convergence, so it may not be effective in finding the global optima, particularly for complex problems. Hence, this paper proposes an intensified SCSO with multiple strategies (IMSCSO). The performance of the IMSCSO algorithm was evaluated on 23 standard test functions and test suites of CEC 2017, CEC 2019, and CEC 2020. Experimental results show that the IMSCSO algorithm performs significantly better than or is on par with other state-of-the-art optimizers. The statistical results obtained from the Wilcoxon signed-rank test and the Friedman test also indicate that the IMSCSO algorithm has a high ability to significantly outperform and rank first among all methods. Moreover, seven typical engineering issues were employed to estimate the efficacy of IMSCSO in optimizing constrained problems. The experimental findings show that the suggested IMSCSO method can efficiently handle real-world application issues.
2 September 2024 in IEEE Access
The Snow Ablation Optimizer (SAO) is a powerful metaheuristic algorithm that emulates the process of snow sublimation and melting. Despite exhibiting competitive performance compared to classical algorithms in initial research, SAO still has certain limitations, including suboptimal convergence accuracy, inadequate population diversity, and premature convergence, particularly when faced with high-dimensional complex challenges. To address these limitations, this paper proposes a refined snow ablation optimizer featuring joint opposite selection and multi-strategy fusion (JMRSAO), which integrates three enhancement strategies. Firstly, a good point set opposition-based learning (GPSOL) approach is utilized to generate a population of high-quality individuals that are evenly distributed. This helps the algorithm converge quickly towards the appropriate search space. Next, in both exploration and exploitation phases, we introduce joint opposite selection (JOS) technology to effectively retain superior individuals for future iterations. This promotes information exchange among diverse individuals and establishes a strong balance between exploring new possibilities and exploiting existing knowledge. Furthermore, we incorporate the Powell mechanism (PM) to enhance local exploration and accelerate convergence towards improved solutions. The performance of JMRSAO is evaluated by comparing it with thirteen other metaheuristic algorithms on three benchmark experiments from IEEE CEC2005, IEEE CEC2017, IEEE CEC2019, and IEEE CEC2022 as well as six engineering problems. Experimental results demonstrate that JMRSAO effectively addresses the limitations of SAO and exhibits exceptional performance in solving various complex problems while showcasing its potential as an effective problem-solving tool.