基于自适应蚁群算法的软硬件划分  

Hardware/software partition using adaptive ant colony algorithm

作　　者：张煜东[1] 吴乐南[1] 韦耿[1] 吴含前[2] 郭永亮[3] 

机构地区：[1]东南大学信息科学与工程学院,南京210096 [2]东南大学软件学院,南京210096 [3]东南大学移动通信国家重点实验室,南京210096

出　　处：《控制与决策》2009年第9期1385-1389,共5页Control and Decision

基　　金：国家自然科学基金项目(60473065;60572063;60872075;60802006);高等学校科技创新工程重大项目培育基金项目(706028);国家863计划项目(2008AA01Z227)

摘　　要：为了更好地解决软硬件双路划分问题,提出一种自适应蚁群算法.基本思想是:对状态转移概率与信息素挥发因子,采取自适应调节策略.保证了算法前期蚁群的随机性较大,可以充分全局搜索;算法后期蚁群的随机性降低,以使算法在较短的时间内收敛.对不同节点的控制数据流图进行仿真实验,表明在同等条件下,相对于改进模拟退火、改进禁忌搜索、改进蚁群算法以及DCG3A方法,所提出算法的命中率与收敛时间结果均更优.节点规模越大,优势尤其明显.In order to solve the hardware/software bi-partitioning problem more efficiently, a novel adaptive ant colony algorithm （AACA） is proposed. The basic idea is to adaptively adjust the state transform probability and the pheromone evaporation factor, which ensures that the randomness of the ant colonies is high enough at the initial for global search and low at the later stage for local search for faster convergence. Experiments synthesize different nodes control data flow graphs, and show that the proposed method has superiority over improved simulated annealing （ISA）, improved tabu search （ITS）, improved ant colony algorithm （IACA）, and DCG3A in terms of global convergence rate and computation time. The more the nodes are, the obvious the superiority is.

关 键 词：嵌入式系统 协同设计 软硬件划分 蚁群算法 

分 类 号：TN911.73[电子电信—通信与信息系统]