MSHHOTSA: A variant of tunicate swarm algorithm combining multi-strategy mechanism and hybrid Harris optimization

Abstract

This paper proposes a novel hybrid algorithm, named Multi-Strategy Hybrid Harris Hawks Tunicate Swarm Optimization Algorithm (MSHHOTSA). The primary objective of MSHHOTSA is to address the limitations of the tunicate swarm algorithm, which include slow optimization speed, low accuracy, and premature convergence when dealing with complex problems. Firstly, inspired by the idea of the neighborhood and thermal distribution map, the hyperbolic tangent domain is introduced to modify the position of new tunicate individuals, which can not only effectively enhance the convergence performance of the algorithm but also ensure that the data generated between the unknown parameters and the old parameters have a similar distribution. Secondly, the nonlinear convergence factor is constructed to replace the original random factor c1 to coordinate the algorithm’s local exploitation and global exploration performance, which effectively improves the ability of the algorithm to escape extreme values and fast convergence. Finally, the swarm update mechanism of the HHO algorithm is introduced into the position update of the TSA algorithm, which further balances the local exploitation and global exploration performance of the MSHHOTSA. The proposed algorithm was evaluated on eight standard benchmark functions, CEC2019 benchmark functions, four engineering design problems, and a PID parameter optimization problem. It was compared with seven recently proposed metaheuristic algorithms, including HHO and TSA. The results were analyzed and discussed using statistical indicators such as mean, standard deviation, Wilcoxon’s rank sum test, and average running time. Experimental results demonstrate that the improved algorithm (MSHHOTSA) exhibits higher local convergence, global exploration, robustness, and universality than BOA, GWO, MVO, HHO, TSA, ASO, and WOA algorithms under the same experimental conditions.

A feature selection method based on the Golden Jackal-Grey Wolf Hybrid Optimization Algorithm

“…Similar to other metaheuristic algorithms, the complexity of the GJO-GWO algorithm is predominantly influenced by three key processes: initialization, fitness evaluation, and individual updating [ 84 , 85 ]. Notably, the complexity of fitness evaluation is intricately dependent on the intricacy of the specific optimization problem under consideration; consequently, we shall refrain from an exhaustive examination of this aspect.…”

A Dynamic Hierarchical Improved Tyrannosaurus Optimization Algorithm with Hybrid Topology Structure

“…These algorithms can compensate for the shortcomings of the original algorithms by integrating new strategies or the advantages of other algorithms to achieve better optimization results. For example, Hu et al suggested an enhanced Jellyfish Optimization Algorithm (EJS) with multiple strategies, and its exceptional efficacy was confirmed through various test functions and practical engineering uses [27]; Li and others suggested a superior dung beetle optimization algorithm (EDBO) suitable for engineering uses, and its efficiency was assessed with the CEC2017 function set, revealing EDBO's robust optimization stability and precision [28]; Zheng and colleagues suggested an advanced multi-strategy African Vulture Optimization Algorithm (EAVOA), which was utilized in multi-layered perceptual classification experiments using XOR and cancer data, demonstrating its superiority over alternative approaches [29]; Jiang and colleagues suggested an enhanced multi-strategy Flamingo Search Algorithm (IFSA), with experimental confirmation of its superior convergence precision and detection ability [30]; Liu and colleagues suggested an algorithm for quilt swarming that merges a multi-strategy approach with hybrid Harris optimization (MSHHOTSA), and confirmed through experiments that MSHHOTSA outperforms BOA, GWO, MVO, HHO, TSA, ASO, and WOA in terms of local convergence, overall exploration capability, resilience, and broad applicability [31]; Song and others introduced an adaptive particle swarm optimization algorithm (APSO/DU) with multiple strategies, opting for an actual MSV portfolio optimization challenge. The findings indicated that the APSO/DU algorithm excels in convergence precision and speed, successfully identifying the stock portfolio with minimal risk while maintaining equivalent returns [32].…”