Property Directed Reachability with Extended Resolution
Andrew Luka, Yakir Vizel
TL;DR
This work introduces PdrER, a model-checking algorithm that extends Property Directed Reachability (Pdr) by employing Extended Resolution ($ER$) to construct shorter proofs of correctness. To manage the challenges of $ER$ in the model-checking setting, the authors implement template-guided extension, a delta-trace re-encoding strategy, and program-specific adjustments such as GeneralizeER, PropagateER, and ImplicationER to handle auxiliary variables. Experimental evaluation on the Hardware Model Checking Competition benchmarks demonstrates that PdrER solves more instances and often with shorter proofs and faster runtimes than Pdr and ABC’s Pdr, including problems that Pdr cannot solve within time limits. The results indicate that strong proof systems like $ER$ can be made practically usable in model checking, suggesting a significant step toward integrating powerful logical frameworks into verification tools.
Abstract
Property Directed Reachability (\textsc{Pdr}), also known as IC3, is a state-of-the-art model checking algorithm widely used for verifying safety properties. While \textsc{Pdr} is effective in finding inductive invariants, its underlying proof system, Resolution, limits its ability to construct short proofs for certain verification problems. This paper introduces \textsc{PdrER}, a novel generalization of \textsc{Pdr} that uses Extended Resolution (ER), a proof system exponentially stronger than Resolution, when constructing a proof of correctness. \PdrEV leverages ER to construct shorter bounded proofs of correctness, enabling it to discover more compact inductive invariants. While \PdrEV is based on \textsc{Pdr}, it includes algorithmic enhancements that had to be made in order to efficiently use ER in the context of model checking. We implemented \textsc{PdrER} in a new open-source verification framework and evaluated it on the Hardware Model Checking Competition benchmarks from 2019, 2020 and 2024. Our experimental evaluation demonstrates that \textsc{PdrER} outperforms \textsc{Pdr}, solving more instances in less time and uniquely solving problems that \textsc{Pdr} cannot solve within a given time limit. We argue that this paper represents a significant step toward making strong proof systems practically usable in model checking.
