摘要
针对面向松耦合可重构系统及大任务集下软硬件划分算法研究的不足,提出了一种基于模拟退火算法和禁忌搜索算法相融合的软硬件划分方法。首先,多次运用与初值无关的模拟退火算法对系统DAG关键路径上的节点进行软硬件划分得到初始划分解;然后,采用禁忌搜索算法在初始划分的基础上进行全局寻优得到近似最优解。与禁忌算法相比,提出的算法划分质量较优,且算法运行时间最多可缩短为原来的53%。
For the lack of research on hardware/software (HW/SW) partitioning algorithms for loosely coupled reconfigurable systems and large task sets, a HW/SW partitioning method is pro- posed based on the simulated annealing (SA) algorithm and tabu search (TS) algorithm. Firstly, an initial partition solution is obtained through repeatedly applying SA algorithm to do the HW/SW partition of targeting nodes on the critical path in the DAG system ; then the TS algorithm is utilized to globally search the approximate optimum solution on the basis of the initial partition. Compared to the TS algorithm, partition quality of the proposed algorithm is relatively superior, and the running time of the proposed algorithm is reduced to at most 53% of the original.
出处
《信息工程大学学报》
2015年第3期364-369,共6页
Journal of Information Engineering University
基金
国家863计划资助项目(2009AA012201)
上海市科委重大科技攻关项目(08DZ501600)
关键词
软硬件划分
关键路径
模拟退火算法
预划分
禁忌搜索算法
hardware/software partitioning
critical path
simulated annealing algorithm
pre-parti-tioning
tabu search algorithm