XOR Satisfiability Solver Module for DPLL Integration

Satisfiability solvers that are based on the Davis-Putnam-Logemann-Loveland (DPLL) algorithm operate on propositional logic formulas in conjunctive normal form (CNF). Despite major improvements in solver technology, using only CNF does not seem to scale well for problem instances involving XOR expressions. We present a decision procedure to determine effectively the satisfiability of XOR clauses and to extract an explanation from conflicting assignments. Our modular proof-of-concept implementation can be integrated tightly into a DPLL-based solver capable of conflict-driven clause learning.