Delete comment from: Computational Complexity
I know this is an old post, but I'm still missing part of the argument, namely the running time bound.
For some sources of randomness the algorithm will finish sooner, for some later -- this makes thinking about s as universal number for every x a bit vague. My understanding is that one needs to bound the expectation of the running time -- is it somehow hidden in the assumption that x is Kolmogorov random string? Or is more work needed?
I also noticed that the paper by Moser and Tardos use somewhat different algorithm, one that avoids recursion.
Is it just a presentation detail, or do the two algorithms have different running times? Or is this the version
from the talk as opposed to the paper?
May 19, 2017, 7:53:29 AM
Posted to A Kolmogorov Complexity Proof of the Lovász Local Lemma