■对算法更新方程的改进是最改进的技术之一。由于缺乏搜索能力,均衡优化器(EO)在求解复杂优化问题时计算复杂度高,可操作性差,本文提出了一种改进的EO,即综合环氧乙烷(MS-EO)更新的多策略。
■首先,在EO中采用了简化的更新策略,以提高可操作性并降低计算复杂度。其次,信息共享策略在简化的EO中使用动态调整策略更新早期迭代阶段的浓度,以形成简化的共享EO(SS-EO)并增强勘探能力。第三,利用迁移策略和黄金分割策略对黄金粒子进行更新,构造黄金SS-EO(GS-EO),提高搜索能力。最后,对后期最差的粒子更新实施精英学习策略,形成MS-EO,增强开发能力。这些策略被嵌入到EO中,通过充分发挥各自的优势,在勘探与开发之间取得平衡。
■来自CEC2013和CEC2017测试集的复杂函数的实验结果表明,MS-EO在搜索能力方面优于EO和相当多最先进的算法,运行速度和可操作性。在多个数据集上进行特征选择的实验结果表明,MS-EO还提供了更多的优势。
UNASSIGNED: Improvement on the updating equation of an algorithm is among the most improving techniques. Due to the lack of search ability, high computational complexity and poor operability of equilibrium optimizer (EO) in solving complex optimization problems, an improved EO is proposed in this article, namely the multi-strategy on updating synthetized EO (MS-EO).
UNASSIGNED: Firstly, a simplified updating strategy is adopted in EO to improve operability and reduce computational complexity. Secondly, an information sharing strategy updates the concentrations in the early iterative stage using a dynamic tuning strategy in the simplified EO to form a simplified sharing EO (SS-EO) and enhance the exploration ability. Thirdly, a migration strategy and a golden section strategy are used for a golden particle updating to construct a Golden SS-EO (GS-EO) and improve the search ability. Finally, an elite learning strategy is implemented for the worst particle updating in the late stage to form MS-EO and strengthen the exploitation ability. The strategies are embedded into EO to balance between exploration and exploitation by giving full play to their respective advantages.
UNASSIGNED: Experimental results on the complex functions from CEC2013 and CEC2017 test sets demonstrate that MS-EO outperforms EO and quite a few state-of-the-art algorithms in search ability, running speed and operability. The experimental results of feature selection on several datasets show that MS-EO also provides more advantages.