具有学习因子的动态搜索烟花算法

发布时间：2018-03-16 06:26

  本文选题：动态搜索烟花算法　切入点：爆炸半径　出处：《计算机科学与探索》2017年03期 　论文类型：期刊论文

【摘要】：采用核心烟花动态爆炸半径策略的动态搜索烟花算法(dynamic search fireworks algorithm,dyn FWA)已被证明是解决优化问题的一个重要算法。然而,dyn FWA的寻优精度低且容易过早地陷入局部最优解。为了改善上述的缺陷,通过嵌入一种利用历史成功信息生成两种不同的学习因子来改进传统的动态搜索烟花算法,称为改进的动态搜索烟花算法(improved dyn FWA,Idyn FWA)。算法中的学习因子充分利用搜索过程中每一代最好的烟花个体信息,使得烟花具有向群体的优良搜索信息学习的能力,并且它的两种不同产生方式有助于平衡算法的局部搜索和全局搜索能力。改进后的算法在CEC2013的28个Benchmark函数上进行测试,实验结果表明Idyn FWA的寻优效果明显优于dyn FWA,并且比粒子群算法SPSO2011和差分演化算法DE/randto-best/1能达到更好的寻优性能。
[Abstract]:The dynamic search fireworks algorithm dyn FWAs (dynamic search fireworks algorithm) has been proved to be an important algorithm to solve the optimization problem. However, the optimization accuracy of FWA is low and it is easy to fall into the local optimal solution prematurely. In order to improve the above deficiencies, The traditional dynamic search fireworks algorithm is improved by embedding two different learning factors using historical success information. The algorithm is called improved dynamic search fireworks algorithm. The learning factors in the algorithm make full use of the best individual information of each generation in the search process, so that fireworks have the ability to learn from the excellent search information of the group. And its two different generation methods are helpful to balance the local search and global search ability of the algorithm. The improved algorithm is tested on 28 Benchmark functions of CEC2013. The experimental results show that the optimization performance of Idyn FWA is better than that of dyn FWA, and it is better than that of particle swarm optimization (SPSO2011) and differential evolution algorithm (DE/randto-best/1).
【作者单位】： 安徽大学计算机科学与技术学院;安徽大学计算智能与信号处理教育部重点实验室;金陵科技学院计算机学院;
【基金】：国家自然科学基金No.61375121 安徽高校省级自然科学研究项目No.KJ2013A009 安徽大学博士启动基金 金科院引进人才科研项目No.jit-rcyj-201505~~
【分类号】：TP18
，

本文编号：1618704