跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-21 12:19 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 05:40 — 2026-11-22 05:40(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-21 12:19最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 1 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-21 12:17(北京时间)
新论文 · v1arXiv:2611.19613

Towards Understanding the Complexity of CAQE: A Proof-Theoretic Analysis of its Core Procedure

Benjamin Böhm, Olaf Beyersdorff

原文摘要

CAQE (Clausal Abstraction for Quantifier Elimination) is currently the most successful algorithmic paradigm for solving Quantified Boolean Formulas (QBF) in practice, clearly dominating recent QBF competitions. While apparently being a very strong solver, not much is known about CAQE in theory. We propose a framework for formalising runs of the basic CAQE algorithm (where failed assumptions are not considered) as proofs in a proof system CAQE$^{\mathsf{Core}}$, which we use to analyse the algorithm proof-theoretically and which provides methods for future work to separate it from optimised versions of CAQE that are used in practice. We show that one can perform strategy extraction on the CAQE$^{\mathsf{Core}}$ proof system in such a way that strategy size serves as a lower bound for basic CAQE runs. Furthermore, we introduce a measure, which we call CAQE width, which not only acts as a lower bound on CAQE$^{\mathsf{Core}}$ proofs, but -- with the number of $\exists\forall$ alternations as exponent -- as an upper bound as well. Using this analysis, we prove that Q-resolution with an NP oracle simulates CAQE$^{\mathsf{Core}}$ on QCNFs of bounded quantifier alternation. The same holds for QCDCL (Quantified Conflict-Driven Clause Learning, formalised as a proof system) on $Σ_3^b$ QCNFs when purely propositional CDCL steps are not counted, after a simple preprocessing step. In this setting, both systems are indeed strictly stronger than CAQE$^{\mathsf{Core}}$, where the separation of QCDCL from CAQE$^{\mathsf{Core}}$ uses cube learning.

cs.LO · cs.CC