...
首页> 外文期刊>The Open Automation and Control Systems Journal >The Application of Computerized Algorithms in the Design Method of Software-hardware Dual-track Partitioning in an Embedded System Abstract
【24h】

The Application of Computerized Algorithms in the Design Method of Software-hardware Dual-track Partitioning in an Embedded System Abstract

机译:计算机算法在嵌入式系统软硬件双轨分区设计方法中的应用

获取原文
           

摘要

It has been proved that the hardware/software partitioning problem is NP-hard. Currently we have tried a varietyof computerized algorithms to resolve it, which can be divided into two major categories: accurate algorithms and heuristicalgorithms. This paper will discuss accurate algorithms and heuristic algorithms respectively. Accurate algorithmstake the example of a greedy algorithm. It abstracts the hardware/software partitioning problem into 01 knapsack modeland obtains the exact optimal solution by the greedy algorithm, while heuristic algorithms use a genetic algorithm as anexample. It converts the hardware/software partitioning problem into a multi-constraint 0-1 knapsack problem and solvesit by employing the genetic algorithm therein. By steps like "variation” and "crossover", this algorithm makes an offspringsolution quickly approach an optimal solution, thereby constructing a near-optimal heuristic solution of the HW/SW partitioningproblem. Experimental results demonstrate that the algorithm proposed in this paper can effectively resolve thehardware/software partitioning problem, have a good global searching capability, and the heuristic algorithm performsfaster than the traditional accurate algorithm, but the heuristic algorithm only acquires a near-optimal solution, which isnot perfect.
机译:已经证明,硬件/软件分区问题是NP难的。目前,我们已经尝试了多种计算机算法来解决该问题,这些算法可以分为两大类:精确算法和启发式算法。本文将分别讨论精确算法和启发式算法。精确算法以贪婪算法为例。该算法将硬件/软件分区问题抽象为01背包模型,并通过贪婪算法获得精确的最优解,而启发式算法以遗传算法为例。它将硬件/软件分区问题转换为多约束0-1背包问题,并通过其中的遗传算法进行求解。通过“变异”和“交叉”等步骤,该算法使后代解快速逼近最优解,从而构造了HW / SW分区问题的近似最优启发式解,实验结果表明,本文提出的算法能够有效地解决。硬件/软件分区问题,具有良好的全局搜索能力,启发式算法比传统的精确算法更快,但是启发式算法只能获得接近最优的解决方案,并不完美。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号