---
格式版本: 2
标题: "DD-suite: A cross-platform package to build Decision Diagrams for optimization purposes"
原文链接: "https://arxiv.org/abs/2608.10957"
发布日期: "2026-08-11"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 11 Aug 2026 14:24:34 UTC (2,023 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-12T11:21:06+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-12T11:16:29+08:00"
入库时间: "2026-08-12T03:21:07.040Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=performance&searchtype=all"
匹配关键词:
  - "performance"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 5
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "arXiv论文，主题为数学优化工具包，与超节点/AI Rack/机柜级AI基础设施完全无关，无技术或商业信号。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-12T11:21:11+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
采集批次: "2026年8月12日2点04分20秒"
采集批次ID: "20260812-020420-289"
去重键: "https://arxiv.org/abs/2608.10957"
---

## Mathematics > Optimization and Control

## Title:DD-suite: A cross-platform package to build Decision Diagrams for optimization purposes

[View PDF](https://arxiv.org/pdf/2608.10957) [HTML (experimental)](https://arxiv.org/html/2608.10957v1)

> Abstract:Decision diagrams (DDs) have become a powerful tool for discrete optimization, supporting a wide range of algorithms that span cut-generation procedures, decomposition methods, and specialized branch-and-bound searches. Despite this growth, their adoption remains limited, partly because most existing DD code is tailored to a specific algorithm or application and is therefore hard to reuse. We introduce DD-suite, a cross-platform, open-source software package for building and manipulating DDs for discrete optimization. DD-suite is available in both Python and C++ through a shared modeling interface, and lets users construct exact, restricted, and relaxed DDs for any discrete optimization problem expressed in recursive form. The package implements the DD reduction procedure, shortest-path routines for obtaining primal and dual bounds, a visualization tool, and an extensive automated test suite. Furthermore, it includes extensive documentation, a support webpage, and ready-to-use examples for four combinatorial problems. Rather than a closed solver, DD-suite is designed as an extensible building block: users can add new construction mechanisms or run custom algorithms on top of the resulting diagram, as we illustrate with a DD-based cutting plane procedure embedded in a state-of-the-art mixed-integer programming solver. Our numerical experiments show that the C++ implementation is 5-6 times faster than the Python one while producing identical diagrams, and remains within a small constant factor of ddo, a specialized Rust framework, confirming that DD-suite combines an accessible, extensible codebase with competitive performance.

| Subjects: | Optimization and Control (math.OC) |
| --- | --- |
| Cite as: | [arXiv:2608.10957](https://arxiv.org/abs/2608.10957) \[math.OC\] |
|  | (or [arXiv:2608.10957v1](https://arxiv.org/abs/2608.10957v1) \[math.OC\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.10957](https://doi.org/10.48550/arXiv.2608.10957) |

## Submission history

From: Margarita Castro \[[view email](https://arxiv.org/show-email/71773bb1/2608.10957)\]  
**\[v1\]** Tue, 11 Aug 2026 14:24:34 UTC (2,023 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.10957) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))