基于GPU加速的细粒度并行人工蜂群算法
A Fine-grained Parallel Artificial Bee Colony Algorithm Based on GPU-Acceleration
-
摘要: 为了改善人工蜂群算法对于大规模数据、高复杂度问题的执行效率,采用开放计算语言(OpenCL)并行编程模型,提出一种基于图形处理器(GPU)加速的并行人工蜂群算法.将每只采蜜蜂映射到OpenCL的一个工作组,跟随蜂采用局部轮盘赌选择,使得人工蜂群算法在GPU中加速执行.实验结果表明,并行人工蜂群算法取得了较好的优化效果,提高了算法的运算速度.Abstract: Artificial Bee Colony(ABC) algorithm still needs plenty of computing time when it processes large-scale data and high complexity problems.In order to solve those problems in practice,we proposed a parallel ABC algorithm based on GPU acceleration with OpenCL parallel programming model.The process of parallel ABC is each employee bee mapping to OpenCL work group and onlooker using local roulette wheel selection,which makes parallel ABC speed up.The simulation results show that parallel ABC algorithm achieves an effective optimization and speeds up its running.
下载:
