| 摘 要: 针对浣熊优化算法(COA)易陷入局部最优、收敛速度快的缺点,本文提出了一种多策略融合的浣熊优化算法(MICOA)。该算法采用自适应适应度距离平衡策略,平衡个体的适应度函数值和个体与最优解的距离,增强了算法跳出局部最优的能力;采用自适应协方差学习策略,COA算法能够在开发阶段充分利用优势种群信息;采用了局部最优扰动方案,有利于帮助算法跳出局部最优。选用CEC2014函数,在收敛精度、收敛速度、统计检验3个方面对改进后算法的优良性进行实验检验。实验结果表明,改进策略有效地提升了原算法的寻优精度与收敛速度。并在工程优化问题上进一步验证策略的实际性。 |
| 关键词: 浣熊优化算法 多策略改进 自适应适应度距离平衡 自适应协方差学习 局部最优扰动 |
|
中图分类号: TP391.7
文献标识码: A
|
|
| Multi-Strategy Improved Coati Optimization Algorithm |
|
ZHAO Hui, DAI Yongqiang
|
(College of Information Science and Technology, Gansu Agricultural University, Lanzhou 730070, China)
416484959@qq.com; daiyq@gsau.edu.cn
|
| Abstract: To address the issues of the Coati Optimization Algorithm (COA) being prone to local optima and exhibiting fast convergence, this paper proposes a Mult-i strategy Improved Coati Optimization Algorithm (MICOA).The approach incorporates an adaptive fitness-distance balance strategy to harmonize individual fitness values and their proximity to the optimal solution, thereby enhancing the algorithm’s ability to escape local optima. An adaptive covariance learning strategy is adopted, enabling COA to fully leverage advantageous population information during the exploitation phase. Additionally, a local optimal perturbation scheme is implemented to facilitate escaping local optima. The CEC2014 benchmark functions are employed to evaluate the enhanced algorithm across three dimensions: convergence accuracy, convergence speed, and statistical testing. Experimental results demonstrate that the proposed strategies significantly improve the optimization precision and convergence speed of the original algorithm. The practicality of these strategies is further validated through engineering optimization problems. |
| Keywords: Coati Optimization Algorithm mult-i strategy improvement adaptive fitness-distance balance adaptive covariance learning local optimal perturbation |
