We revisit two well-studied problems,
Bounded Degree Vertex Deletion
and
Defective Coloring
, where the input is a graph
G
and a target degree
Δ
and we are asked either 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.
We revisit the parameterization by treewidth, as well as several related parameters and present a more fine-grained picture of the complexity of both problems. In particular:
•
Both problems admit straightforward DP algorithms with table sizes (
Δ
+ 2)
tw
and (
χ
d
(
Δ
+ 1))
tw
respectively, where tw is the input graph’s treewidth and
χ
d
the number of available colors. We show that, under the SETH, both algorithms are essentially optimal, for any non-trivial fixed values of
Δ
,
χ
d
, even if we replace treewidth by pathwidth. Along the way, we obtain an algorithm for
Defective Coloring
with complexity quasi-linear in the table size, thus settling the complexity of both problems for treewidth and pathwidth.
•
Given that the standard DP algorithm is optimal for treewidth and pathwidth, we then go on to consider the more restricted parameter tree-depth. Here, previously known lower bounds imply that, under the ETH,
Bounded Vertex Degree Deletion
and
Defective Coloring
cannot be solved in time
\(n^{o(\sqrt [4]{\mathrm{td}})} \)
and
\(n^{o(\sqrt {\mathrm{td}})} \)
respectively, leaving some hope that a qualitatively faster algorithm than the one for treewidth may be possible. We close this gap by showing that neither problem can be solved in time
n
o
(td)
, under the ETH, by employing a recursive low tree-depth construction that may be of independent interest.
•
Finally, we consider a structural parameter that is known to be restrictive enough to render both problems FPT: vertex cover. For both problems the best known algorithm in this setting has a super-exponential dependence of the form
\(\mathrm{vc}^{\mathcal {O}(\mathrm{vc})} \)
. We show that this is optimal, as an algorithm with dependence of the form vc
o
(vc)
would violate the ETH. Our proof relies on a new application of the technique of
d
-detecting families introduced by Bonamy et al. [ToCT 2019].
Our results, although mostly negative in nature, paint a clear picture regarding the complexity of both problems in the landscape of parameterized complexity, since in all cases we provide essentially matching upper and lower bounds.