首页> 外文会议>International conference on knowledge science, engineering and management >An Efficient Approach for Computing Conflict Sets Combining Failure Probability with SAT
【24h】

An Efficient Approach for Computing Conflict Sets Combining Failure Probability with SAT

机译:一种将失败概率与SAT相结合的计算冲突集的有效方法

获取原文

摘要

According to the topological structure of circuit and the characteristics of the enumeration tree, with the in-depth study of the reversed depth set enumeration tree method to compute conflict sets (CSRDSE), an efficient approach for counting conflict sets combining failure probability with SAT (CSCFPS) is proposed. Firstly, this paper presents the concept of fault output component set, possible fault component set, non-fault component set, explained fault component set, failure probability and non-conflict set theorem. Secondly, the OrderedCS-FP algorithm is came up. This algorithm calculates an ordered component set where the components are organized in descending order of failure probability. Finally, this paper puts forward the CSCFPS algorithm. In this algorithm, the SE-Tree is generated on the basis of the ordered component set, so minimal conflict set can be visited as early as possible to avoid the access to redundant nodes. On the grounds of the non-conflict set theorem, the sub-tree corresponding to the non-fault component set is deleted, decreasing the traversal to these nodes. Thus, the number of calling the SAT solver is greatly lessened. The experimental results show that the efficiency of this method is significantly improved.
机译:根据电路的拓扑结构和枚举树的特性,通过对反向深度集枚举树方法计算冲突集(CSRDSE)的深入研究,一种有效的将失败概率与SAT相结合的冲突集的计算方法( CSCFPS)。首先,本文提出了故障输出分量集,可能的故障分量集,非故障分量集,解释的故障分量集,故障概率和非冲突定理的概念。其次,提出了OrderedCS-FP算法。该算法计算有序组件集,其中组件以失败概率的降序排列。最后提出了CSCFPS算法。在该算法中,SE-Tree是基于有序组件集生成的,因此可以尽早访问最小冲突集,以避免访问冗余节点。基于非冲突集定理,删除了与非故障组件集相对应的子树,从而减少了对这些节点的遍历。因此,大大减少了调用SAT解算器的次数。实验结果表明,该方法的效率明显提高。

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