Core Concepts
The core message of this paper is that the standard dynamic programming algorithms for Bounded Degree Vertex Deletion and Defective Coloring are essentially optimal when parameterized by treewidth, pathwidth, tree-depth, and vertex cover.
Abstract
The paper revisits two well-studied problems, Bounded Degree Vertex Deletion and Defective Coloring, where the input is a graph G and a target degree Δ, and the goal is to edit or partition the graph so that the maximum degree becomes bounded by Δ. Both problems are known to be parameterized intractable for the most well-known structural parameters, such as treewidth.
The authors present a more fine-grained picture of the complexity of both problems:
For treewidth and pathwidth, they show that the standard dynamic programming algorithms are essentially optimal. Specifically, they prove that under the SETH, no algorithm can solve Bounded Degree Vertex Deletion and Defective Coloring in time (Δ+2-ε)^pw and (χd(Δ+1)-ε)^pw respectively, where pw is the pathwidth of the input graph, and χd is the number of available colors for Defective Coloring.
For tree-depth, the authors show that under the ETH, neither problem can be solved in time no(td), where td is the tree-depth of the input graph. This implies that the treewidth-based algorithm remains (qualitatively) optimal even in this more restricted case.
For vertex cover, the authors prove that the best known algorithms, which have a super-exponential dependence on the vertex cover, are optimal under the ETH. This is shown by a new application of the technique of d-detecting families.
Overall, the results paint a clear picture of the structurally parameterized complexity of these two problems, indicating that the standard dynamic programming algorithms are optimal in a multitude of restricted cases.