A backtracking survey propagation algorithm for K-satisfiability

TL;DR

This paper introduces a backtracking enhancement to the survey propagation algorithm, significantly improving its effectiveness in solving challenging K-satisfiability problems near the phase transition.

Contribution

The paper presents a novel backtracking version of survey propagation, advancing the algorithm's capability to handle difficult instances of K-satisfiability.

Findings

Backtracking improves survey propagation's success rate on hard problems

Enhanced algorithm better solves instances near the sat-unsat transition

Demonstrates the effectiveness of backtracking in combinatorial algorithms

Abstract

In this paper we present a backtracking version of the survey propagation algorithm. We show that the introduction of the simplest form of backtracking greatly improves the ability of the original survey propagation algorithm in solving difficult random problems near the sat-unsat transition.