一种嵌入式系统软硬件划分算法  

Simulation Research on HW-SW Partitioning of Embedded System Based on Ant Colony Algorithm

作　　者：李正民[1] 郭金金[1] 吕莹莹[1] 

机构地区：[1]郑州大学信息工程学院,河南郑州450001

出　　处：《计算机仿真》2011年第10期204-207,235,共5页Computer Simulation

摘　　要：软硬件划分是嵌入式系统软硬件协同设计中的一个关键问题。传统划分算法具有局部最优,收敛速度慢的缺陷。为使组成系统性能达到最优化,提出一种新的嵌入式系统软硬件划分算法。先采用嵌入式系统转化成有向无环图,可将嵌入式系统软硬件划分问题转换成一个多条件约束问题,用蚂蚁放置于有向无环图顶点上,对系统软硬件的划分准确率作为蚁群算法优化目标,通过蚁群算法搜索最优目标函数值,有效避免传统划分算法搜索陷入局部最小,大幅度降低搜索时间。实验结果表明,采用蚁群算法能够高效、快速获得准确地划分结果,为嵌入式系统设计提供了依据。Hardware/software partition is a key problem in collaborative design of embedded system.Because the traditional partition methods have the defects of local optimal and slow convergence speed,the article put forward a hardware/software partition method of embedded system based on ant colony algorithm.This method used acyclic graph to mode embedded system,and the hardware/software partition problems were converted to a multi-conditions problem.Then the ant colony was put in system nodes,and the partitioning rate of the system hardware and software was taken as the optimal objective.Through ant colony algorithm,optimal objective function values were searched,avoiding the local minima and reduceing the searching time significantly.Compared with genetic algorithm and simulated annealing algorithm,the experimental results showed that the ant colony algorithm can faster search speed and get better hardwares/software partition results,and the nodes scale is larger.

关 键 词：嵌入式系统 软硬件划分 蚁群算法 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]