基于OpenMP的并行粒子群优化算法研究  

Research on parallel particle swarm optimization algorithm based on OpenMP

作　　者：康军广[1] 段国林[1] 王金敏[2] 田永军[1] 

机构地区：[1]河北工业大学机械工程学院,天津300130 [2]天津职业技术师范大学机械工程学院,天津300222

出　　处：《河北工业大学学报》2015年第2期34-37,共4页Journal of Hebei University of Technology

基　　金：国家自然科学基金(60975046)

摘　　要：针对现有粒子群优化算法多采用串行方式执行且运行效率较低的问题,提出一种基于Open MP技术的并行粒子群优化算法.该算法以多核硬件平台为基础,利用粒子群算法搜索速度快,易于并行等特点,引入Open MP技术,通过将该并行算法应用于布局问题求解并与串行算法相比较,测试结果表明,该并行算法与串行算法结果一致,能够充分利用多核CPU的计算资源,运行效率得到明显提高.Concerning the low performance and executing in sequential way of most particle swarm optimization algorithms, a parallel algorithm based on OpenMP was proposed. By introducing OpenMP technology, the algorithm with the advantages of searching fast and easy to be parallelized of PSO is based on multi-core hardware platform. The parallel particle swarm algorithm is tested by solving the rectangular layout instance and is compared with the sequential way. The experimental results we got are the same as by choosing the sequential way, also the results show that the proposed algorithm has improved the efficiency in solving the rectangular layout problem by making full use of the multicore computing resources.

关 键 词：OPENMP 并行 粒子群 多核CPU 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]