Title:Lopsidependency in the Moser-Tardos framework: Beyond the Lopsided Lovasz Local Lemma

Abstract:The Lopsided Lovász Local Lemma (LLLL) is a powerful probabilistic principle which has been used in a variety of combinatorial constructions. While originally a general statement about probability spaces, it has recently been transformed into a variety of polynomial-time algorithms. The resampling algorithm of Moser & Tardos (2010) is the most well-known example of this. A variety of criteria have been shown for the LLLL; the strongest possible criterion was shown by Shearer, and other criteria which are easier to use computationally have been shown by Bissacot et al (2011), Pegden (2014), Kolipaka & Szegedy (2011), and Kolipaka, Szegedy, Xu (2012).
We show a new criterion for the Moser-Tardos algorithm to converge. This criterion is stronger than the LLLL criterion; this is possible because it does not apply in the same generality as the original LLLL; yet, it is strong enough to cover many applications of the LLLL in combinatorics. We show a variety of new bounds and algorithms. A noteworthy application is for $k$-SAT, with bounded occurrences of variables. As shown in Gebauer, Szábo, and Tardos (2011), a $k$-SAT instance in which every variable appears $L \leq \frac{2^{k+1}}{e (k+1)}$ times, is satisfiable. Although this bound is asymptotically tight (in $k$), we improve it to $L \leq \frac{2^{k+1} (1 - 1/k)^k}{k-1} - \frac{2}{k}$ which can be significantly stronger when $k$ is small.
We introduce a new parallel algorithm for the LLLL. While Moser & Tardos described a simple parallel algorithm for the Lovász Local Lemma, and described a simple sequential algorithm for a form of the Lopsided Lemma, they were not able to combine the two. Our new algorithm applies in nearly all settings in which the sequential algorithm works --- this includes settings covered by our new stronger LLLL criterion.