Birdie Blog

Thinking will not overcome fear but action will.

Towards Real-World Routing with Neural Combinatorial Optimization

真实世界、非对称

Towards Real-World Routing with Neural Combinatorial Optimization Jiwoo Son, Zhikai Zhao, Federico Berto, Chuanbo Hua, Zhiguang Cao, Changhyun Kwon, and Jinkyoo Park ICLR2026,rating:4648 https:/...

Towards Efficient Constraint Handling in Neural Solvers for Routing Problems

复杂约束、用微调后的解来监督训练

Towards Efficient Constraint Handling in Neural Solvers for Routing Problems Jieyi Bi, Zhiguang Cao, Jianan Zhou, Wen Song, Yaoxin Wu, Jie Zhang, Yining Ma, and Cathy Wu ICLR2026,rating:6486 htt...

RADAR - Learning to Route with Asymmetry-aware Distance Representations

非对称问题、SVD+Sinkhorn

RADAR: Learning to Route with Asymmetry-aware Distance Representations Hang Yi, Ziwei Huang, Yining Ma, and Zhiguang Cao ICLR2026,rating:4464 https://openreview.net/forum?id=lWdxX5s9T1 引入 针对非对...

Chain-of-Context Learning - Dynamic Constraint Understanding for Multi-Task VRPs

多任务、动态信息和静态信息关联

Chain-of-Context Learning: Dynamic Constraint Understanding for Multi-Task VRPs Shuangchun Gui, Suyu Liu, Xuehe Wang, and Zhiguang Cao ICLR2026,rating:44666 https://openreview.net/forum?id=AhE6a...

An Agentic Framework with LLMs for Solving Complex Vehicle Routing Problems

LLM agent CO

An Agentic Framework with LLMs for Solving Complex Vehicle Routing Problems Ni Zhang, Zhiguang Cao, Jianan Zhou, Cong Zhang, and Yew-Soon Ong ICLR2026,rating:4664 https://openreview.net/forum?id...

Efficiently Solving the Practical Vehicle Routing Problem - A Novel Joint Learning Approach

看看KDD的NCO

Efficiently Solving the Practical Vehicle Routing Problem: A Novel Joint Learning Approach 蚂蚁集团 KDD2020 贡献 本文提出一种基于图卷积网络(GCN)的模型,以节点特征(坐标与需求)和边特征(节点间真实距离)为输入并进行嵌入。我们设计了两个独立的解码器分别对两类嵌入进行解码。一个解码器...

Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation

LLM合成数据解决train和test数据分布不一致

Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation zhiguang cao课题组 AAAI2026 研究背景和问题 近年来,基于深度强化学习的神经求解器在合成数据(如均匀分布的TSP或CVRP)上表现优异,但在真实世界基准...

ICLR 2026 review阶段,LLM AHD合集四

LLM AHD合集四

ICLR 2026 review阶段,LLM AHD合集四 AutoEP: LLMs-Driven Automation of Hyperparameter Evolution for Metaheuristic Algorithms https://openreview.net/forum?id=hit3hGBheP 国防科大 rating:8666, Accept, Oral ...

ICLR 2026 review阶段,LLM AHD合集三

LLM AHD合集三

ICLR 2026 review阶段,LLM AHD合集三 Cognitively Inspired Reflective Evolution: Interactive Multi-Turn LLM–EA Synthesis of Heuristics for Combinatorial Optimization https://openreview.net/forum?id=31VTD...

ICLR 2026 review阶段,LLM AHD合集二

LLM AHD合集二

ICLR 2026 review阶段,LLM AHD合集二 Fusing LLMs with Scientific Literature for Heuristic Discovery https://openreview.net/forum?id=lwqeXDYKWJ rating:4444, Reject 核心思想 让大语言模型(LLM)在进化算法中“查文献”,从而突破自身知识...