Research map / Label-efficient and robust learning / Recognition and long-tail
RADAR: Learning to Route with Asymmetry-aware Distance Representations
Hang Yi, Ziwei Huang, Yining Ma, Zhiguang Cao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Sinkhorn normalization to attention (ablation) | RADAR · ATSP1000 | optimality gap: 22.89 -> 4.13 | Improved |
| add SVD initialization (ablation) | RADAR · ATSP1000 | optimality gap: 7.24 -> 4.13 | Improved |
| add Sinkhorn alone to baseline (ablation) | RADAR · ATSP1000 | optimality gap: 38.64 -> 7.24 | Improved |
| add SVD alone to baseline (ablation) | RADAR · ATSP1000 | optimality gap: 38.64 -> 22.89 | Improved |
| add SVD initialization (ablation) | RADAR · ATSP100 | optimality gap: 1.19 -> 0.72 | Improved |
| RADAR instead of ReLD | ReLD · ATSP1000 | optimality gap: 31.67 -> 4.13 | Improved |
| RADAR instead of ReLD | ReLD · ACVRP1000 | optimality gap: 94.45 -> 0.96 | Improved |
| RADAR in multi-task framework instead of RF-NN | RouteFinder · 16 asymmetric VRP variants | average gap (%): 1.9875 -> 1.3331 | Improved |
| RADAR instead of RRNCO on real-world ACVRPTW | RRNCO · Real-world ACVRPTW (in-distribution) | gap (%): 3.93 -> 2.71 | Improved |
| remove coordinates from RRNCO (ablation) | RRNCO · ATSP100 in-distribution | gap: 2.60% -> 3.22% | Hurt |
| remove coordinates from RADAR (ablation) | RADAR · ATSP100 in-distribution | gap: 1.52% -> 1.49% | No clear effect |
| add POMO coordinate augmentation to RADAR (ablation) | RADAR · ATSP100 in-distribution | gap: 1.52% -> 0.74% | Improved |
Closest papers by meaning
- SPACE: Unifying Symmetric and Asymmetric Routing Problems for Generalist Neural Solver (NeurIPS 2026 · similarity 0.92)
- A Mixed-Curvature based Pre-training Paradigm for Multi-Task Vehicle Routing Solver (ICML 2025 · similarity 0.88)
- Multi-Task Vehicle Routing Solver via Mixture of Specialized Experts under State-Decomposable MDP (NeurIPS 2025 · similarity 0.87)
- MVMoE: Multi-Task Vehicle Routing Solver with Mixture-of-Experts (ICML 2024 · similarity 0.86)
- DPN: Decoupling Partition and Navigation for Neural Solvers of Min-max Vehicle Routing Problems (ICML 2024 · similarity 0.86)
- SHIELD: Multi-task Multi-distribution Vehicle Routing Solver with Sparsity and Hierarchy (ICML 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
