Abstract: To address the issues of uneven population distribution during initialization, susceptibility to local optima in complex search spaces, and insufficient optimization capability in the later exploitation phase of the Hippopotamus Optimization Algorithm (HO), an improved HO incorporating coupled disturbance and covariance evolution is proposed. First, in the initialization phase, Latin hypercube sampling is employed to optimize the initial population distribution,enhancing population diversity and global coverage. Second, in the exploration phase, a coupled interference strategy is introduced to strengthen the algorithm's global search capability in high-dimensional spaces, reducing the risk of premature convergence to local optima. Finally, in the exploitation phase, a covariance matrix adaptation evolutionary strategy is incorporated to adaptively adjust the covariance matrix and step-size parameters, enhancing the algorithm's dynamic adaptability to search space structures and improving both convergence speed and accuracy in late-stage local search. Experimental evaluations were conducted on 12 multi-modal benchmark functions from the CEC2017 benchmark functions, comparing the proposed HO against 8 state-of-the-art swarm intelligence algorithms and performing statistical analysis. Statistical analysis of the results demonstrates that the proposed algorithm achieves improvements of 32.46% in convergence accuracy and reduces standard deviation by 60.1% compared to the HO, indicating its superior convergence performance and enhanced stability, thereby validating the effectiveness of the proposed improvement strategies.

Key words: hippopotamus optimization algorithm, latin hypercube sampling, coupled interference strategy, covariance matrix adaptation evolution strategy, swarm intelligence algorithm

摘要： 针对河马优化算法在初始化阶段种群分布不均、复杂搜索空间中易陷入局部最优以及后期开发阶段寻优能力不足的问题，提出基于耦合干扰与协方差进化改进的河马优化算法。首先，在初始化阶段，采用拉丁超立方采样优化初始种群分布，提高种群多样性和全局覆盖范围。其次，在探索阶段，融合耦合干扰策略强化算法在高维空间中的全局搜索能力，降低陷入局部最优的风险。最后，在开发阶段，引入协方差矩阵适应性进化策略，自适应调整协方差矩阵及步长参数，增强算法对搜索空间结构的动态适应能力，提升迭代后期局部搜索的收敛速度与精度。在CEC2017多类型的12个基准测试函数上，与8种主流群智能算法进行实验对比及统计分析。结果表明，相较于河马优化算法收敛精度与标准差分别提升了32.46%和60.1%，展现了改进算法较好的收敛性及稳定性，验证了改进策略的有效性。

关键词: 河马优化算法, 拉丁超立方采样, 耦合干扰策略, 协方差矩阵适应性进化策略, 群智能算法