Given once, 2009
On the Infeasibility of Polynomial Kernelization
A 2009 tour of why some hard problems just will not shrink down to polynomial size
An intermediate thesis defence, for researchers and faculty.
Abstract
Draft, not yet checked by meA professor wants 6 popular guests out of 25 at a party: does a small dominating set exist? In my master's thesis talk, I use examples like this one to introduce kernelization (polynomial-time preprocessing that shrinks an instance to a size bounded by a parameter) and ask when polynomial-size kernels cannot exist. I cover the composition and distillation framework, polynomial parameter transformations, and worked examples (weighted satisfiability, Disjoint Factors, and Vertex-Disjoint Cycles). I also look at ways around the barrier, such as the many polynomial kernels for out-branchings with many leaves, and at some open problems.
The slides
As a PDF
The viewer needs JavaScript; the PDF below works without it.