A generalization of Nemhauser and Trotter's local optimization theorem

  • Authors:

  • Michael R. Fellows;Jiong Guo;Hannes Moser;Rolf Niedermeier

  • Affiliations:

  • School of Engineering and Information Technology, Charles Darwin University, Darwin, Northern Territory 0909, Australia;Universität des Saarlandes, Campus E 1.7, D-66123 Saarbrücken, Germany;Institut für Informatik, Friedrich-Schiller-Universität Jena, Ernst-Abbe-Platz 2, D-07743 Jena, Germany;Institut für Softwaretechnik und Theoretische Informatik, TU Berlin, D-10587 Berlin, Germany

  • Venue:
  • Journal of Computer and System Sciences
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Nemhauser-Trotter local optimization theorem applies to the NP-hard Vertex Cover problem and has applications in approximation as well as parameterized algorithmics. We generalize Nemhauser and Trotter@?s result to vertex deletion problems, introducing a novel algorithmic strategy based on purely combinatorial arguments (not referring to linear programming as the Nemhauser-Trotter result originally did). The essence of our strategy can be understood as a doubly iterative process of cutting away ''easy parts'' of the input instance, finally leaving a ''hard core'' whose size is (almost) linearly related to the cardinality of the solution set. We exhibit our approach using a generalization of Vertex Cover, called Bounded-Degree Vertex Deletion. For some fixed d=0, Bounded-Degree Vertex Deletion asks to delete at most k vertices from a graph in order to transform it into a graph with maximum vertex degree at most d. Vertex Cover is the special case of d=0. Our generalization of the Nemhauser-Trotter-Theorem implies that Bounded-Degree Vertex Deletion, parameterized by k, admits an O(k)-vertex problem kernel for d=0, an O(k^1^+^@e)-vertex problem kernel for d=2. Finally, we provide a W[2]-completeness result for Bounded-Degree Vertex Deletion in case of unbounded d-values.