Research map / Reinforcement learning / Visual and representation RL
USPR: Learning a Unified Solver for Profiled Routing
Chuanbo Hua, Federico Berto, Zhikai Zhao, Jiwoo Son, Changhyun Kwon, Jinkyoo Park
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 profile-aware score reshaping (PSR) (ablation) | USPR · PVRP N=100 | avg gap (%): 4.72 -> 4.42 | Improved |
| add score reshaping (SR, distance only) on top of PSR (ablation) | USPR · PVRP N=100 | avg gap (%): 5.12 -> 4.72 | Improved |
| add multi-head profiled attention (MHPA) on top of PSR and SR (ablation) | USPR · PVRP N=100 | avg gap (%): 6.72 -> 5.12 | Improved |
| single unified model instead of 10 separate CAMP models | USPR · PVRP training cost | training time and parameters: 4.6 days -> 11 hours; 17.6M -> 1.5M parameters | Improved |
| USPR versus CAMP, greedy decoding | USPR · PVRP in-distribution | avg gap (%): 10.50 -> 9.20 | Improved |
| USPR versus CAMP on large-scale N=1000, sampling | USPR · PVRP large-scale | avg gap (%): 23.44 -> 14.15 | Improved |
| USPR versus CAMP on PVRPLib large instances | USPR · PVRPLib Set X (700-1K) | gap to BKS (%): 25.77 -> 15.21 | Improved |
Closest papers by meaning
- Rethinking Light Decoder-based Solvers for Vehicle Routing Problems (ICLR 2025 · similarity 0.86)
- Multi-Task Vehicle Routing Solver via Mixture of Specialized Experts under State-Decomposable MDP (NeurIPS 2025 · 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)
- MVMoE: Multi-Task Vehicle Routing Solver with Mixture-of-Experts (ICML 2024 · similarity 0.85)
- Memory-Enhanced Neural Solvers for Routing Problems (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.
