English

Simultaneously Satisfying Linear Equations Over $\mathbb{F}_2$: MaxLin2 and Max-$r$-Lin2 Parameterized Above Average

Data Structures and Algorithms 2011-05-17 v3 Computational Complexity

Abstract

In the parameterized problem \textsc{MaxLin2-AA}[kk], we are given a system with variables x1,...,xnx_1,...,x_n consisting of equations of the form iIxi=b\prod_{i \in I}x_i = b, where xi,b{1,1}x_i,b \in \{-1, 1\} and I[n],I\subseteq [n], each equation has a positive integral weight, and we are to decide whether it is possible to simultaneously satisfy equations of total weight at least W/2+kW/2+k, where WW is the total weight of all equations and kk is the parameter (if k=0k=0, the possibility is assured). We show that \textsc{MaxLin2-AA}[kk] has a kernel with at most O(k2logk)O(k^2\log k) variables and can be solved in time 2O(klogk)(nm)O(1)2^{O(k\log k)}(nm)^{O(1)}. This solves an open problem of Mahajan et al. (2006). The problem \textsc{Max-rr-Lin2-AA}[k,rk,r] is the same as \textsc{MaxLin2-AA}[kk] with two differences: each equation has at most rr variables and rr is the second parameter. We prove a theorem on \textsc{Max-rr-Lin2-AA}[k,rk,r] which implies that \textsc{Max-rr-Lin2-AA}[k,rk,r] has a kernel with at most (2k1)r(2k-1)r variables improving a number of results including one by Kim and Williams (2010). The theorem also implies a lower bound on the maximum of a function f: {1,1}nRf:\ \{-1,1\}^n \rightarrow \mathbb{R} of degree rr. We show applicability of the lower bound by giving a new proof of the Edwards-Erd{\H o}s bound (each connected graph on nn vertices and mm edges has a bipartite subgraph with at least m/2+(n1)/4m/2 + (n-1)/4 edges) and obtaining a generalization.

Keywords

Cite

@article{arxiv.1104.1135,
  title  = {Simultaneously Satisfying Linear Equations Over $\mathbb{F}_2$: MaxLin2 and Max-$r$-Lin2 Parameterized Above Average},
  author = {R. Crowston and M. Fellows and G. Gutin and M. Jones and F. Rosamond and S. Thomasse and A. Yeo},
  journal= {arXiv preprint arXiv:1104.1135},
  year   = {2011}
}
R2 v1 2026-06-21T17:50:23.901Z