面向单目标优化的集成粒子群算法

发布时间：2018-09-09 12:07

【摘要】：串行粒子群算法广泛应用于多个领域,出现了多个变种,但解决不同种类的优化问题时性能有差异。为提高串行粒子群算法对各种优化问题的适应能力,提出一种集成粒子群优化算法。新算法使用Matlab的单程序多数据并行结构发挥单节点多核计算能力,通过设置外部档案分享不同粒子群的全局最佳位置,促进不同串行粒子群算法之间的信息交流,综合利用不同串行粒子群算法在解决不同类型优化问题的优势。在广泛使用的测试函数集上开展仿真实验,结果验证了新算法的有效性,与多个知名的串行粒子群算法相比,新算法在寻优性能上优势明显。新算法不仅能够提高粒子群算法的适应能力,而且,所采用的算法框架也适应于其他群智能算法,改善了算法的性能。
[Abstract]:Serial particle swarm optimization algorithm is widely used in many fields, and there are many varieties, but the performance is different when solving different kinds of optimization problems. In order to improve the adaptability of serial particle swarm optimization to various optimization problems, an integrated particle swarm optimization algorithm is proposed. The new algorithm uses Matlab's single-program multi-data parallel structure to play a single point and multi-core computing power, by setting the external files to share the global optimal location of different particle swarm optimization, to promote the exchange of information between different serial particle swarm optimization algorithms. Comprehensive use of different serial particle swarm optimization algorithm in solving different types of optimization problems. Simulation experiments are carried out on the widely used test function set, and the results show that the new algorithm is effective. Compared with several well-known serial particle swarm optimization algorithms, the new algorithm has obvious advantages in optimization performance. The new algorithm can not only improve the adaptability of particle swarm optimization algorithm, but also adapt to other swarm intelligence algorithms, and improve the performance of the algorithm.
【作者单位】：河南工程学院计算机学院;
【基金】：国家自然科学基金青年项目(61301232,61501174) 河南省高等学校重点科研项目(17A520026) 河南工程学院博士基金(D2012016)~~
【分类号】：TP18

【参考文献】