摘要
为改善群居蜘蛛优化算法中存在的收敛速度较慢和收敛精度低的问题,提出基于混沌映射与高斯变异的群居蜘蛛优化算法。首先,在种群初始化过程中采用混沌映射反向学习策略;其次,蜘蛛位置更新过程中引入自适应权重和动态概率因子;最后,在种群完成交配操作后,针对蜘蛛群中最优个体位置进行高斯变异扰动。通过实验证明:改进后的算法在测试函数上能够做到快速收敛,同时函数的最终收敛精度得到明显提高。通过30维与50维的实验结果可以看出,改进后的算法对于不同维度的函数都有着较好的收敛精度。
To improve the problems of slow convergence rate and low convergence accuracy existing in the gregarious spider optimization algorithm,a social spider optimization algorithm based on chaos mapping and gaussian mutation is proposed.Firstly,the chaotic mapping reverse learning strategy is adopted during population initialization;Secondly,adaptive weights and dynamic probability factors are introduced in the spider position update;Finally,gaussian variation is disturbed for the optimal individual position in the spider group.Experiments show that the improved algorithm can converge quickly in the test function,and the final convergence accuracy of the function is significantly improved.The experimental results of 30 dimensions and 50 dimensions show that the improved algorithm has a good convergence accuracy for the functions of different dimensions.
作者
叶坤涛
郜海毅
李晟
YE Kuntao;GAO Haiyi;LI Sheng(College of Science,Jiangxi University of Science and Technology,Ganzhou Jiangxi 341000)
出处
《软件》
2022年第5期1-7,共7页
Software
基金
江西省教育厅科技项目(GJJ170526)。
关键词
群居蜘蛛优化算法
高斯变异
混沌映射
动态概率因子
收敛精度
social spider optimization
gaussian mutation
chaos mapping
dynamic probability factor
convergence accuracy
