基于多核PC的人工蜂群并行算法的研究与实现  

Research and Implementation of Parallel Random Perturbation Artificial Bee Colony Algorithm Based on Multi-core PC

作　　者：李双[1] 李文敬[1] 杨文[1] 周海燕[1] 

机构地区：[1]广西师范学院计算机与信息工程学院,广西南宁530023

出　　处：《微电子学与计算机》2012年第9期63-66,70,共5页Microelectronics & Computer

基　　金：国家自然科学基金项目(61163012);广西自然科学基金项目(0991105)

摘　　要：针对人工蜂群算法搜索时间长、收敛速度缓慢、搜索后期极易陷入局部最优等问题,提出了一种人工蜂群并行算法(PRP-ABC).首先引入随机摄动因子和全局最优解两个策略得到串行式RP-ABC算法;然后应用OpenMP并行技术,采用主从式调度方式对串行算法并行改写;最后在多核PC上将三种算法在复杂函数优化问题上测试比较.结果表明,该算法有效提高了收敛速度和优化性能,尤其适用于高维复杂函数的优化问题.As considering the problem of search time long, and easily into local optimal solution in later stage ofevolution in Artificial Bee Colony （ABC） algorithm, a Parallel Random Perturbation Artificial Bee Colony （PRP-ABC） algorithm is proposed out. Firstly, introduced two concepts here which Random Perturbation factor and global optimal solution, get a serial type RP-ABC algorithm; Secondly, using OpenMP parallel rewrite it, this paper using master-slave scheduling model~ Finally, tested the three algorithm about complex function optimization problem under the multi-core PC environment. The result shows that the algorithm convergence speed and the optimization performance had been improved, especially suitable for the high dimension of complex function optimization problems.

关 键 词：人工蜂群算法 并行算法 随机摄动因子 多核 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]