一种混合正余弦算法的改进海马优化器

A modified sea-horse optimizer with a hybrid sine cosine algorithm

海马优化器是一种模拟海马移动、捕食和繁殖行为的元启发式算法,具有控制参数少、易于部署等优点。为了克服海马优化器在某些复杂情况下收敛速度较慢且易陷入局部最优的不足,本文提出一种混合正余弦算法的改进海马优化器,记作HSCASHO。首先,通过调整参数对海马优化器的数学模型进行修正,更好地平衡全局探索和局部开采;其次,引入一个基于适应度值的权重,并为原始正余弦算法设计一个新的搜索公式,加快收敛速度;最后,增强变异过程种群多样性,有助于算法跳出局部最优。综合两种算法的优点,HSCASHO在前3/4次迭代使用进行参数调整的海马优化器,后1/4次迭代使用更新搜索公式的正余弦算法。将HSCASHO算法与4种算法在10个基准函数上进行对比实验,表明该算法明显优于其他算法。

The sea-horse optimizer(SHO)is a meta-heuristic algorithm that simulates the movement,predation,and reproduction behavior of seahorses.It offers the benefits of minimal control parameters and effortless deployment.To address the limitations of the SHO in complex scenarios,such as slow convergence speed and easy falling into the local optimal,a modified algorithm of hybrid sine cosine algorithm and sea-horse optimizer,called HSCASHO,is proposed in this paper.Firstly,the mathematical model of the SHO is modified by parameter adjustment to better balance global exploration and local exploitation;Secondly,a weight based on fitness value is introduced,and a new search formula is designed for the original sine cosine algorithm to accelerate the convergence speed;Finally,the process of variation enhances population diversity and helps the algorithm to escape the local optimal.Combining the advantages of both algorithms,HSCASHO uses a parameter-tuned SHO for the first 3/4 iterations and a sine cosine algorithm that updates the search formula for the last 1/4 iterations.Experiments comparing the HSCASHO algorithm with the four algorithms on 10 benchmark functions show that the algorithm significantly outperforms the other algorithms.