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

+1171新增论文
+938其中 AI 相关
+130已有论文修订
+94新增 AI 导读

arXiv 最近完成扫描:2026-11-19 06:59 · 全站最近入库:2026-11-19 17:46各信源进度 →

统计口径与数据新鲜度

统计 2026-11-18 17:55 — 2026-11-19 17:55(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:2026-11-19 06:59最近发现新增版本:2026-11-19 17:46库内论文最新更新时间:2026-11-19 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。

27,410 已收录论文22,788 AI 相关8,343 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 2026-11-19 06:57(北京时间)
新论文 · v1arXiv:2611.20310

Strengthening Compact Vehicle Routing Formulations through Arc-Flow Aggregation and Refinement

Julian Yarkony, Willem-Jan van Hoeve, Anthony Karahalios, Martin Karyamsetty, Soheyl Khalilpourazari, Louis-Martin Rousseau, Xiaohang Zhu

原文摘要

Arc-flow formulations can provide strong relaxations for mixed-integer program formulations of vehicle routing problems, but the associated networks may be prohibitively large. We study the setting in which multiple arc flow relaxations may be applied, and introduce a dual-guided aggregation and refinement framework that constructs substantially smaller arc-flow formulations while recovering the LP bound of the original networks. Dual information is used both to identify states that can be aggregated without weakening the bound and to guide node disaggregation. We show that the resulting procedure converges finitely to the arc-flow bound of the full arc-flow formulation. The framework retains the original decision variables, allowing cutting planes to be added directly in the original variable space. Moreover, it produces an explicit MILP formulation without relying on column generation or pricing to determine the arc-flow relaxation bound. We apply our approach to the vehicle routing problem with time windows for which we define two arc flow relaxations, as well as two families of cutting planes over the original variable space. Computational experiments show that strong bounds can be obtained using substantially aggregated networks. The resulting root bounds are competitive with those obtained by the branch-and-price solver RouteOpt, while retaining a formulation that can be solved directly by a general-purpose MILP solver.

math.OC · cs.DM