针对霜冰优化算法的搜索策略单一化,算法后期搜索开发能力有限,导致算法稳定性不足,提出了一种基于t分布扰动因子和随机差分变异算子的改进策略。在霜冰优化算法的硬刺穿透机制的基础上,引入了t分布扰动因子,局部范围内扩大算法的搜索范...针对霜冰优化算法的搜索策略单一化,算法后期搜索开发能力有限,导致算法稳定性不足,提出了一种基于t分布扰动因子和随机差分变异算子的改进策略。在霜冰优化算法的硬刺穿透机制的基础上,引入了t分布扰动因子,局部范围内扩大算法的搜索范围,试图在最优位置周围探索更优的位置。在算法迭代完成后,利用随机差分变异算子对最新更新的粒子位置进行突变,得到更优的粒子。通过测试集CEC2017和CEC2022,与同类算法进行对比实验,发现改进后的霜冰优化算法搜索能力更强,稳定性更好。同时进行了Wilcoxon符号秩检验,验证了算法的显著性差异。利用经典测试函数,从最优值迭代曲线、平均适应度值、第一维度变化趋势和种群历史位置4个维度展开,对算法特征进行了分析。最后应用改进后的霜冰优化算法优化PID(proportional integral derivative)参数进行仿真实验,验证了算法的有效性和适用性。展开更多
针对算术优化算法(Arithmetic optimization algorithm, AOA)易陷入局部最优、收敛速度慢等问题,提出一种基于T分布透镜成像反向学习策略的算术优化算法(TOBLAOA),设计动态平衡机制调节全局探索强度与局部开采深度的配比,从而突破早熟...针对算术优化算法(Arithmetic optimization algorithm, AOA)易陷入局部最优、收敛速度慢等问题,提出一种基于T分布透镜成像反向学习策略的算术优化算法(TOBLAOA),设计动态平衡机制调节全局探索强度与局部开采深度的配比,从而突破早熟收敛瓶颈并提高解精度稳定性。对2005、2017、2015测试集中的部分基准函数(共15个)进行仿真实验,首先引入4种不同改进策略的改进透镜成像反向学习策略的算术优化算法(OBLAOA)算法进行比较,再与哈里斯鹰优化算法、雷电附着优化算法等6个优化算法进行了实验结果对比和差异显著性Wilcoxon秩和检验,结果表明改进后的算术优化算法在求解精度、收敛速度上均有显著提升。最后将算法应用于工程优化中的常见的压力容器设计、三杆桁架设计、齿轮系设计中,进一步验证此算法的有效性。To address issues such as the Arithmetic Optimization Algorithm (AOA) being prone to local optima and having slow convergence speed, this paper proposes a T-distribution lens imaging opposition-based learning arithmetic optimization algorithm (TOBLAOA). The algorithm designs a dynamic balance mechanism to regulate the ratio between global exploration intensity and local exploitation depth, thereby breaking through the bottleneck of premature convergence and improving the stability of solution accuracy. Simulation experiments were conducted on selected benchmark functions (totaling 15) from the 2005, 2017, and 2015 test sets. The study first compares four improved OBLAOA algorithms with different enhancement strategies, and then performs experimental result comparisons and Wilcoxon rank-sum significance tests with six optimization algorithms including Harris Hawk Optimization and Lightning Attachment Procedure Optimization. The results demonstrate that the enhanced arithmetic optimization algorithm achieves significant improvements in both solution accuracy and convergence speed. Finally, the algorithm was applied to common engineering optimization problems such as pressure vessel design, three-bar truss design, and gear train design, further verifying its effectiveness.展开更多
文摘针对霜冰优化算法的搜索策略单一化,算法后期搜索开发能力有限,导致算法稳定性不足,提出了一种基于t分布扰动因子和随机差分变异算子的改进策略。在霜冰优化算法的硬刺穿透机制的基础上,引入了t分布扰动因子,局部范围内扩大算法的搜索范围,试图在最优位置周围探索更优的位置。在算法迭代完成后,利用随机差分变异算子对最新更新的粒子位置进行突变,得到更优的粒子。通过测试集CEC2017和CEC2022,与同类算法进行对比实验,发现改进后的霜冰优化算法搜索能力更强,稳定性更好。同时进行了Wilcoxon符号秩检验,验证了算法的显著性差异。利用经典测试函数,从最优值迭代曲线、平均适应度值、第一维度变化趋势和种群历史位置4个维度展开,对算法特征进行了分析。最后应用改进后的霜冰优化算法优化PID(proportional integral derivative)参数进行仿真实验,验证了算法的有效性和适用性。
文摘针对算术优化算法(Arithmetic optimization algorithm, AOA)易陷入局部最优、收敛速度慢等问题,提出一种基于T分布透镜成像反向学习策略的算术优化算法(TOBLAOA),设计动态平衡机制调节全局探索强度与局部开采深度的配比,从而突破早熟收敛瓶颈并提高解精度稳定性。对2005、2017、2015测试集中的部分基准函数(共15个)进行仿真实验,首先引入4种不同改进策略的改进透镜成像反向学习策略的算术优化算法(OBLAOA)算法进行比较,再与哈里斯鹰优化算法、雷电附着优化算法等6个优化算法进行了实验结果对比和差异显著性Wilcoxon秩和检验,结果表明改进后的算术优化算法在求解精度、收敛速度上均有显著提升。最后将算法应用于工程优化中的常见的压力容器设计、三杆桁架设计、齿轮系设计中,进一步验证此算法的有效性。To address issues such as the Arithmetic Optimization Algorithm (AOA) being prone to local optima and having slow convergence speed, this paper proposes a T-distribution lens imaging opposition-based learning arithmetic optimization algorithm (TOBLAOA). The algorithm designs a dynamic balance mechanism to regulate the ratio between global exploration intensity and local exploitation depth, thereby breaking through the bottleneck of premature convergence and improving the stability of solution accuracy. Simulation experiments were conducted on selected benchmark functions (totaling 15) from the 2005, 2017, and 2015 test sets. The study first compares four improved OBLAOA algorithms with different enhancement strategies, and then performs experimental result comparisons and Wilcoxon rank-sum significance tests with six optimization algorithms including Harris Hawk Optimization and Lightning Attachment Procedure Optimization. The results demonstrate that the enhanced arithmetic optimization algorithm achieves significant improvements in both solution accuracy and convergence speed. Finally, the algorithm was applied to common engineering optimization problems such as pressure vessel design, three-bar truss design, and gear train design, further verifying its effectiveness.
