Competitive Sorter-based Encoding of PB-Constraints into SAT

14 pages•Published: March 15, 2019

Abstract

A Pseudo-Boolean (PB) constraint is a linear inequality constraint over Boolean variables. A popular idea to solve PB-constraints is to transform them to CNFs (via BDDs, adders and sorting networks [5, 11]) and process them using – increasingly improving – state-of-the-art SAT-solvers. Recent research have favored the approach that uses Binary Decision Diagrams (BDDs), which is evidenced by several new constructions and optimizations [2, 21]. We show that encodings based on comparator networks can still be very competitive. We present a system description of a PB-solver based on MiniSat+ [11] which we extended by adding a new construction of selection network called 4-Way Merge Selection Network, with a few optimizations based on other solvers. Experiments show that on many instances of popular benchmarks our technique outperforms other state-of-the-art PB-solvers.

Keyphrases: comparator network, constraints solver, odd even network, pseudo boolean, sat, selection network

In: Daniel Le Berre and Matti Järvisalo (editors). Proceedings of Pragmatics of SAT 2015 and 2018, vol 59, pages 65-78.

Links:	https://easychair.org/publications/paper/tsHw
	https://doi.org/10.29007/hh3v