首页> 外文期刊>IFAC PapersOnLine >DSM-based variable ordering heuristic for reduced computational effort of symbolic supervisor synthesis ?
【24h】

DSM-based variable ordering heuristic for reduced computational effort of symbolic supervisor synthesis ?

机译:基于DSM的变量排序启发式,减少符号管理综合的计算工作

获取原文
           

摘要

We consider the influence that the variable order of Binary Decision Diagrams (BDDs) has on the computational effort that is required for symbolic supervisor synthesis. In recent research it has been shown that improving the variable order can result in a substantial decrease of synthesis effort. We propose the combined use of the Dependency Structure Matrices (DSMs) and the matrix reordering heuristics Cuthill-McKee and Sloan to minimize the Weighted Event Span (WES) of the variable order. This is done by placing variables that often appear together in transition relations near each other. By performing benchmark experiments, we measure the reduction in synthesis effort by utilizing a variable order with minimized WES. The experiments show that our approach is competitive in reducing computational effort compared to FORCE, a state of practice variable ordering heuristic. Moreover, the best improvements in effort reduction are shown for the most computationally demanding models tested.
机译:我们考虑了二进制决策图(BDD)对符号主管合成所需的计算工作的可变顺序的影响。在最近的研究中,已经表明,改善可变顺序可能导致合成努力的显着降低。我们提出了依赖结构矩阵(DSM)和矩阵重新排序启发式Cuthill-Mckee和Sloan的汇总使用,以最小化可变顺序的加权事件跨度(WES)。这是通过放置在彼此附近的过渡关系中一起出现在一起的变量来完成的。通过执行基准实验,我们通过利用具有最小化WE的可变顺序来测量综合工作的减少。实验表明,与武力相比,我们的方法在减少计算努力方面是竞争力,实践状态可变排序启发式。此外,为测试所测试的大多数计算要求苛刻的模型显示了减少的最佳改进。

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