计算机科学与探索 ›› 2023, Vol. 17 ›› Issue (5): 1075-1088.DOI: 10.3778/j.issn.1673-9418.2108008

• 理论·算法 • 上一篇    下一篇

改进平衡优化器算法在约束优化问题中的应用

李守玉,何庆,陈俊   

  1. 贵州大学 大数据与信息工程学院,贵阳 550025
  • 出版日期:2023-05-01 发布日期:2023-05-01

Application of Improved Equilibrium Optimizer Algorithm to Constrained Optimization Problems

LI Shouyu, HE Qing, CHEN Jun   

  1. College of Big Data and Information Engineering, Guizhou University, Guiyang 550025, China
  • Online:2023-05-01 Published:2023-05-01

摘要: 针对平衡优化器算法存在种群勘探与开发难以平衡、粒子进化信息不足、容易出现早熟现象等问题,提出改进的平衡优化器算法。首先,根据算法优化进行的迭代阶段采用正弦池策略动态地平衡勘探与开发能力,迭代前期通过固定角频率的正弦递减进行大范围的全局勘探,扩大算法探索搜索空间中未知区域,增强发现潜藏优质粒子的能力;迭代后期通过变化角频率的正弦递增进行局部开发使勘探与开发自适应平衡,提高算法优化精度。其次,自适应优先引力策略引入当前最优粒子信息克服粒子进化信息匮乏的问题,然后通过融入均匀分布和贝塔分布共同作用丰富种群粒子进化信息,提高粒子之间的信息交换速率,增强粒子逃离局部最优的能力,达到引导种群向全局最优方向快速收敛目的。最后,使用16个基准测试函数、CEC2017函数集、Friedman检验、Wilcoxon秩和检验以及2个现实中的工程约束优化问题测试所提算法的寻优能力。实验结果表明,相比其他新提出的智能算法,所提算法具有更高的优化精度和更快的收敛速度。

关键词: 平衡优化器算法, 勘探与开发, 约束工程优化问题

Abstract: Aiming at the problems of the equilibrium optimizer algorithm, such as difficult balance between population exploration and exploitation, insufficient information of particle evolution and prematurity, an improved equilibrium optimizer algorithm is proposed. Firstly, in the iterative stage optimized by the algorithm, the sinusoidal pool strategy is used to balance the exploration and development capabilities dynamically. In the early stage of iteration, a large range of global exploration is carried out through the sinusoidal decrease of fixed angular frequency to expand the algorithm to explore unknown areas in the search space and enhance the ability of discovering potential high-quality particles. At the end of iteration, local exploitation is carried out by sinusoidal increase of changing angular frequency to balance exploration and exploitation adaptively and improve the optimization accuracy of the algorithm. Secondly, the adaptive priority gravity strategy introduces the current optimal particle information to overcome the lack of evolution information, enriches the evolution information of the population particles by incorporating the uniform distribution and beta distribution together, improves the information exchange rate between particles, enhances the escape of the particles from the local area, and achieves the goal of guiding the population to converge rapidly towards the global optimum. Finally, 16 benchmark functions, CEC2017 functions, Friedman test, Wilcoxon rank sum test and two real-world engineering constraint optimization problems are used to test the optimization ability of the proposed algorithm. Experimental results show that the proposed algorithm has higher optimization accuracy and faster convergence speed compared with other new proposed intelligent algorithms.

Key words: equilibrium optimizer algorithm, exploration and exploitation, constrained engineering optimization problem