Research map / Architectures and efficiency / Attention and transformers
Learning to Iteratively Solve Routing Problems with Dual-Aspect Collaborative Transformer
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use dual-aspect (separate node and position) embeddings instead of single-aspect (ablation) | DACT · TSP100 T=1k | gap (%): 3.49 -> 1.62 | Helped |
| use dual-aspect embeddings instead of single-aspect (ablation) | DACT · TSP100 T=5k | gap (%): 1.55 -> 0.61 | Helped |
| use dual-aspect embeddings instead of single-aspect (ablation) | DACT · TSP50 T=1k | gap (%): 0.35 -> 0.14 | Helped |
| use cyclic positional encoding instead of absolute PE (ablation) | DACT · TSP100 (trained on TSP50) | gap (%): 7.93 -> 2.98 | Helped |
| use cyclic positional encoding instead of absolute PE (ablation) | DACT · TSP20 (trained on TSP50) | gap (%): 0.21 -> 0.10 | Helped |
| add curriculum learning strategy to PPO training (ablation) | DACT · CVRP20 training | sample efficiency / variance | Helped |
| increase inference steps from T=1k to T=10k | DACT · TSP100 | gap (%): 1.62 -> 0.37 | Helped |
| add x4 coordinate augmentation at inference | DACT · TSP100 | gap (%): 0.37 -> 0.09 | Helped |
Closest papers by meaning
- DPN: Decoupling Partition and Navigation for Neural Solvers of Min-max Vehicle Routing Problems (ICML 2024 · similarity 0.86)
- INViT: A Generalizable Routing Problem Solver with Invariant Nested View Transformer (ICML 2024 · similarity 0.86)
- Rethinking Light Decoder-based Solvers for Vehicle Routing Problems (ICLR 2025 · similarity 0.86)
- CaDA: Cross-Problem Routing Solver with Constraint-Aware Dual-Attention (ICML 2025 · similarity 0.86)
- SPACE: Unifying Symmetric and Asymmetric Routing Problems for Generalist Neural Solver (NeurIPS 2026 · similarity 0.86)
- Multi-Task Vehicle Routing Solver via Mixture of Specialized Experts under State-Decomposable MDP (NeurIPS 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.