Practically Feasible Proof Logging for Pseudo-Boolean Optimization
Paper i proceeding, 2025
Certifying solvers have long been standard for decision problems in Boolean satisfiability (SAT), allowing for proof logging and checking with very limited overhead, but developing similar tools for combinatorial optimization has remained a challenge. A recent promising approach covering a wide range of solving paradigms is pseudo-Boolean proof logging, but this has mostly consisted of proof-of-concept works far from delivering the performance required for real-world deployment. In this work, we present an efficient toolchain based on VeriPB and CakePB for formally verified pseudo-Boolean optimization. We implement proof logging for the full range of techniques in the state-of-the-art solvers RoundingSat and Sat4j, including core-guided search and linear programming integration with Farkas certificates and cut generation. Our experimental evaluation shows that proof logging and checking performance in this much more expressive paradigm is now quite close to the level of SAT solving, and hence is clearly practically feasible.
0-1 integer linear programming
certifying algorithms
proof logging
certification
pseudo-Boolean solving
combinatorial optimization
Författare
Wietze Koops
Köpenhamns universitet
Lunds universitet
Daniel Le Berre
Université d'Artois
Magnus Myreen
Göteborgs universitet
Chalmers, Data- och informationsteknik, Formella metoder
Jakob Nordström
Köpenhamns universitet
Lunds universitet
Andy Oertel
Köpenhamns universitet
Lunds universitet
Yong Kiam Tan
Nanyang Technological University
Agency for Science, Technology and Research (A*STAR)
Marc Vinyals
University of Auckland
Leibniz International Proceedings in Informatics, LIPIcs
18688969 (ISSN)
Vol. 340 219783959773805 (ISBN)
31st International Conference on Principles and Practice of Constraint Programming, CP 2025
Glasgow, United Kingdom,
Glasgow, United Kingdom,
De nästa 700 verifierade kompilatorerna
Vetenskapsrådet (VR) (2021-05165), 2022-01-01 -- 2025-12-31.
Ämneskategorier (SSIF 2025)
Datavetenskap (datalogi)
DOI
10.4230/LIPIcs.CP.2025.21
Relaterade dataset
Practically Feasible Proof Logging for Pseudo-Boolean Optimization: Experimental Data [dataset]
DOI: 10.5281/zenodo.15628603