arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1102.3813cs.DS
收藏

大规模超图对偶化的有效算法

Efficient Algorithms for Dualizing Large-Scale Hypergraphs

Keisuke Murakami, Takeaki Uno

摘要:

超图${\cal F}$是定义在顶点集$V$上的集合族。${\cal F}$的对偶是$V$的最小子集$H$的集合,使得对于任何$F\in {\cal F}$都有$F\cap H \ne \emptyset$。对偶的计算等价于许多问题,如子集族的最小碰集计数、最小集覆盖计数、超图断面计数等。尽管已经提出了许多算法来解决这个问题,但据我们所知,没有一个算法可以处理大规模输入和大量输出的最小碰集。本文的重点是开发时间和空间有效的算法来解决这个问题。我们提出了两个新的算法,新的搜索方法,新的修剪方法,和快速技术的最小检查。计算实验表明,我们的算法是相当快的,即使是大规模的输入,现有的算法不终止在一个实际的时间。

英文摘要:

A hypergraph ${\cal F}$ is a set family defined on vertex set $V$. The dual of ${\cal F}$ is the set of minimal subsets $H$ of $V$ such that $F\cap H \ne \emptyset$ for any $F\in {\cal F}$. The computation of the dual is equivalent to many problems, such as minimal hitting set enumeration of a subset family, minimal set cover enumeration, and the enumeration of hypergraph transversals. Although many algorithms have been proposed for solving the problem, to the best of our knowledge, none of them can work on large-scale input with a large number of output minimal hitting sets. This paper focuses on developing time- and space-efficient algorithms for solving the problem. We propose two new algorithms with new search methods, new pruning methods, and fast techniques for the minimality check. The computational experiments show that our algorithms are quite fast even for large-scale input for which existing algorithms do not terminate in a practical time.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑