Research map / Reinforcement learning / Visual and representation RL
A Hierarchical Reinforcement Learning Based Optimization Framework for Large-scale Dynamic Pickup and Delivery Problems
Yi Ma, Xiaotian Hao, Jianye HAO, Jiawen Lu, Xing Liu, Xialiang Tong, Mingxuan Yuan, Zhigang Li, Jie Tang, Zhaopeng Meng
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 |
|---|---|---|---|
| replace fixed 10-min interval partition with upper-level RL agent (ablation) | hierarchical RL with lower-level RL · DPDP 300-1 | objective (avg distance, km): 126.95 -> 122.42 | Improved |
| replace fixed interval partition with upper-level RL agent (ablation) | hierarchical RL · DPDP 300-2 | objective (avg distance, km): 154.39 -> 142.33 | Improved |
| replace random operator selection with lower-level RL policy (ablation) | 10min-Interval · DPDP 300-1 | objective (avg distance, km): 139.16 -> 126.95 | Improved |
| replace random operator selection with lower-level RL policy (ablation) | 10min-Interval · DPDP 300-2 | objective (avg distance, km): 166.31 -> 154.39 | Improved |
| train on 300-order data instead of 1000-order data (ablation) | hierarchical RL · DPDP 1000-1 | objective (avg distance, km): 159.18 -> 170.78 | Hurt |
| use full hierarchical RL instead of Greedy | Greedy · DPDP 300-1 | objective (avg distance, km): 147.78 -> 122.42 | Improved |
| use full hierarchical RL instead of 1order-Interval E2ERL | E2ERL · DPDP 300-1 | objective (avg distance, km): 158.39 -> 122.42 | Improved |
| use lower-level RL instead of ALNS on static PDP | ALNS · Static PDP 50-1 | solve time: 10728.34s -> 1459.23s | Improved |
| use lower-level RL instead of ALNS on static PDP | ALNS · Static PDP 50-1 | objective (avg distance, km): 96.98 -> 82.43 | Improved |
Closest papers by meaning
- One Step is Enough: Multi-Agent Reinforcement Learning Based on One-Step Policy Optimization for Order Dispatch on Ride-Sharing Platforms (NeurIPS 2026 · similarity 0.84)
- NondBREM: Nondeterministic Offline Reinforcement Learning for Large-Scale Order Dispatching (AAAI 2024 · similarity 0.84)
- Hierarchical Optimization via LLM-Guided Objective Evolution for Mobility-on-Demand Systems (NeurIPS 2025 · similarity 0.83)
- GARLIC: GPT-Augmented Reinforcement Learning with Intelligent Control for Vehicle Dispatching (AAAI 2025 · similarity 0.83)
- Distributionally Robust Multi-Agent Reinforcement Learning for Dynamic Chute Mapping (ICML 2025 · similarity 0.81)
- DPN: Decoupling Partition and Navigation for Neural Solvers of Min-max Vehicle Routing Problems (ICML 2024 · similarity 0.81)
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.
