<p>A multi-labelled tree (or MUL-tree) is a rooted tree leaf-labelled by a set of labels, where each label may appear more than once in the tree. We consider the MUL-tree Set Pruning for Consistency problem (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\textsf{MULSETPC}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPC</mi> </math></EquationSource> </InlineEquation>), which takes as input a set of MUL-trees and asks whether there exists a perfect pruning of each MUL-tree that results in a consistent set of single-labelled trees. <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\textsf{MULSETPC}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPC</mi> </math></EquationSource> </InlineEquation> was proven to be NP-complete by Gascon et al. when the MUL-trees are binary, each leaf label is used at most three times, and the number of MUL-trees is unbounded. To determine the computational complexity of the problem when the number of MUL-trees is constant was left as an open problem. Here, we resolve this question by proving a much stronger result, namely that <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\textsf{MULSETPC}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPC</mi> </math></EquationSource> </InlineEquation> is NP-complete even when there are only two MUL-trees, every leaf label is used at most twice, and either every MUL-tree is binary or every MUL-tree has constant height. Furthermore, we introduce an extension of <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\textsf{MULSETPC}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPC</mi> </math></EquationSource> </InlineEquation> that we call <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\textsf{MULSETPComp}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPComp</mi> </math></EquationSource> </InlineEquation>, which replaces the notion of consistency with compatibility. We prove that <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(\textsf{MULSETPComp}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPComp</mi> </math></EquationSource> </InlineEquation> is NP-complete even in two highly restricted cases – when there are only two MUL-trees, every MUL-tree has constant height and either every leaf label is used at most thrice or the instance contains a single-labelled tree containing all leaf labels. Finally, we present a polynomial-time algorithm for a special case of <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(\textsf{MULSETPC}\)</EquationSource> <EquationSource Format="MATHML"><math> <mi mathvariant="sans-serif">MULSETPC</mi> </math></EquationSource> </InlineEquation> with a constant number of binary MUL-trees where every leaf label appears exactly once in at least one MUL-tree.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

MUL-Tree Pruning for Consistency and Compatibility

  • Christopher Hampson,
  • Daniel J. Harvey,
  • Costas S. Iliopoulos,
  • Jesper Jansson,
  • Zara Lim,
  • Wing-Kin Sung

摘要

A multi-labelled tree (or MUL-tree) is a rooted tree leaf-labelled by a set of labels, where each label may appear more than once in the tree. We consider the MUL-tree Set Pruning for Consistency problem ( \(\textsf{MULSETPC}\) MULSETPC ), which takes as input a set of MUL-trees and asks whether there exists a perfect pruning of each MUL-tree that results in a consistent set of single-labelled trees. \(\textsf{MULSETPC}\) MULSETPC was proven to be NP-complete by Gascon et al. when the MUL-trees are binary, each leaf label is used at most three times, and the number of MUL-trees is unbounded. To determine the computational complexity of the problem when the number of MUL-trees is constant was left as an open problem. Here, we resolve this question by proving a much stronger result, namely that \(\textsf{MULSETPC}\) MULSETPC is NP-complete even when there are only two MUL-trees, every leaf label is used at most twice, and either every MUL-tree is binary or every MUL-tree has constant height. Furthermore, we introduce an extension of \(\textsf{MULSETPC}\) MULSETPC that we call \(\textsf{MULSETPComp}\) MULSETPComp , which replaces the notion of consistency with compatibility. We prove that \(\textsf{MULSETPComp}\) MULSETPComp is NP-complete even in two highly restricted cases – when there are only two MUL-trees, every MUL-tree has constant height and either every leaf label is used at most thrice or the instance contains a single-labelled tree containing all leaf labels. Finally, we present a polynomial-time algorithm for a special case of \(\textsf{MULSETPC}\) MULSETPC with a constant number of binary MUL-trees where every leaf label appears exactly once in at least one MUL-tree.