广义中心混合蛙跳算法

PDF下载HTML

Title:

Shuffled frog-leaping algorithm based on the general center

作者:

赵嘉, 吕莉, 樊棠怀

南昌工程学院 信息工程学院, 江西 南昌 330099

Author(s):

ZHAO Jia, LYU Li, FAN Tanghuai

School of Information Engineering, Nanchang Institute of Technology, Nanchang 330099, China

关键词:

蛙跳算法; 混合蛙跳算法; 广义中心; 蛙跳规则; 群智能算法

Keywords:

frog-leaping algorithm; shuffled frog leaping algorithm (SFLA); general center; frog leaping rule; swarm intelligence algorithms

分类号:

TP301

DOI:

10.3969/j.issn.1673-4785.201405070

文献标志码:

A

摘要:

为解决标准混合蛙跳算法族群之间信息共享能力差的问题,加强族群内蛙的学习能力,利用各族群最优蛙位置的平均中心,构造一个与各族群最优蛙都有关联的虚拟广义中心蛙,提出广义中心混合蛙跳算法.该算法在进化过程中,首先蛙群最优蛙在原有位置及广义中心蛙的位置上进行“贪婪”选择,选择最好位置作为新的族群最优蛙位置;其次将广义中心蛙的优势运用于蛙跳规则中,在标准混合蛙跳算法的蛙跳规则中加入族群最差蛙向广义中心蛙学习的能力.将本文算法与不同维度下的标准混合蛙跳算法及新近提出的知名群智能算法进行比较,实验结果表明,本文算法在解的精度、收敛速度及解的稳定性等方面具有更优的性能.

Abstract:

In this paper, a shuffled frog-leaping algorithm based on general center (GC-SFLA) is proposed to solve the problem of weak information sharing between memeplexes in the shuffled frog leaping algorithm (SFLA) to enhance the learning ability and use the average center of optimal frog. The proposed GC-SFLA generates a virtual general center frog from the optimal frog of each memeplex. Firstly, the optimal frog selects the best location among the original location and general center greedily as new location of new memeplex. After that, the advantage of general center frog is applied to the frog-leaping rule, which enable the worst frog to learn from the general center frog. Experiments are conducted on a set of swarm intelligence algorithms to verify that the new approach outperforms SFLA in different dimensions. The experiment results present promising performance of the GC-SFLA on convergence velocity, precision and stability of solution.