花朵授粉算法的优化

发布时间：2018-03-06 12:19

本文选题：花朵授粉　切入点：高斯变异　出处：《计算机工程与设计》2017年06期 　论文类型：期刊论文

【摘要】：针对花朵授粉算法(FPA)寻优过程中局部深度搜索能力弱、易陷入局部最优、后期收敛速度慢的问题,提出一种基于自适应高斯变异的混合蛙跳花朵授粉算法(AGM-SFLFPA)。借鉴混合蛙跳算法(SFLA)思想,对种群个体按照适应度值进行排序、分组并更新各分组中最差个体的位置,增强算法的局部深度搜索能力并增加种群多样性;通过公示牌动态监测算法是否陷入局部最优,当陷入时,将自动对全局最优个体执行高斯变异操作,提高个体跳出局部最优的能力、增强种群多样性、加快收敛速度。通过6个典型的标准测试函数从4个方面验证该算法的有效性,验证结果表明,AGMSFLFPA具有更好的稳定性和可靠性、更快的收敛速度及更高的寻优精度,适用于高维复杂多极值函数求解问题。
[Abstract]:In order to solve the problem of weak local depth search ability, easy to fall into local optimum and slow convergence rate in the process of flower pollination algorithm (FPA), A hybrid frog-leapfrog flower pollination algorithm based on adaptive Gao Si mutation is proposed. Using the idea of hybrid leapfrog algorithm (SFLAs) for reference, the population individuals are sorted according to fitness, and the worst individual positions in each grouping are updated. Enhance the local depth search ability of the algorithm and increase the diversity of the population, and dynamically monitor whether the algorithm falls into the local optimum by using the bulletin board, when the algorithm is trapped, it will automatically carry out Gao Si mutation operation on the globally optimal individual. This paper improves the ability of individual to jump out of local optimum, enhances population diversity, and accelerates convergence speed. The validity of the algorithm is verified from four aspects by six typical standard test functions, and the results show that AGMSFLFPA has better stability and reliability. Faster convergence speed and higher optimization accuracy are suitable for solving complex multiextremum function problems in high dimension.
【作者单位】：中国矿业大学计算机科学与技术学院;常熟理工学院计算机科学与工程学院;山东理工大学农业工程与食品科学学院;
【基金】：国家自然科学基金项目(61173130)
【分类号】：TP18

【相似文献】