The Effect of The Migration Time on The Parallel Particle Swarm Optimization Algorithm

2020 28th Signal Processing and Communications Applications Conference (SIU) Pub Date : 2020-10-05 DOI:10.1109/SIU49456.2020.9302476

Zýnnet Duygu Akþehýr, Selçuk Aslan

引用次数: 0

Abstract

The particle swarm optimization algorithm is a population-based algorithm used to solve optimization problems. In recent years, a parallel version of the PSO algorithm has proposed to solve complex application problems, and the situation has become the focus of researchers. Communication in this approach is ensured by the migration of particles between sub-population. In the parallel version of the PSO algorithm, parameters such as migration strategy, migration period, the topological structure must be determined. In this study, it was analyzed in which periods it would be more appropriate to perform migration procedures. In this context, the maximum number of iterations of 10000 is divided into three period intervals: 1000-3000, 3000-6000 and 6000-9000. Then, for cases where the number of threads is 2 and 4, the migration process is analyzed over these periods.

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

迁移时间对并行粒子群优化算法的影响

粒子群优化算法是一种基于种群的求解优化问题的算法。近年来，人们提出了一种并行版本的粒子群算法来解决复杂的应用问题，并成为研究人员关注的焦点。这种方法中的通信是通过粒子在子种群之间的迁移来保证的。在并行版本的粒子群算法中，必须确定迁移策略、迁移周期、拓扑结构等参数。在本研究中，分析了在哪个时期执行迁移程序更合适。在这种情况下，将10000的最大迭代次数分为1000 ~ 3000、3000 ~ 6000和6000 ~ 9000三个时间段。然后，对于线程数为2和4的情况，在这些时间段内分析迁移过程。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

2020 28th Signal Processing and Communications Applications Conference (SIU)

自引率

0.00%

发文量

期刊最新文献

Skin Lesion Classification With Deep CNN Ensembles Design of a New System for Upper Extremity Movement Ability Assessment Stock Market Prediction with Stacked Autoencoder Based Feature Reduction Segmentation networks reinforced with attribute profiles for large scale land-cover map production Encoded Deep Features for Visual Place Recognition