Research map / Learning theory and algorithms / Prediction and decision losses
Fast Bellman Updates for Wasserstein Distributionally Robust MDPs
Zhuodong Yu, Ling Dai, Shaohang Xu, Siyang Gao, Chin Pang Ho
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 |
|---|---|---|---|
| Use nested-bisection Fast algorithm instead of Gurobi | Wasserstein DRMDP Bellman update (q=1) · Random DRMDP N=S=A=40 | runtime (s): 22.6013 -> 0.5112 | Improved |
| Use Fast algorithm instead of first-order method FOM (3 its) | Wasserstein DRMDP Bellman update (q=1) · Random DRMDP N=S=A=40 | runtime (s): 29.22 -> 0.5112 | Improved |
| Use Fast algorithm instead of Gurobi at smallest size | Wasserstein DRMDP Bellman update (q=1) · Random DRMDP N=S=A=10 | runtime (s): 0.039 -> 0.0253 | Improved |
| Use Fast algorithm instead of FOM at size 10 | Wasserstein DRMDP Bellman update (q=1) · Random DRMDP N=S=A=10 | runtime (s): 0.644 -> 0.0253 | Improved |
| Scale problem size from 10 to 40 for Fast | Fast algorithm (q=1) · Random DRMDP N=S=A | runtime (s): 0.0253 -> 0.5112 | Hurt |
| Scale problem size from 10 to 40 for Gurobi | Gurobi (q=1) · Random DRMDP N=S=A | runtime (s): 0.039 -> 22.6013 | Hurt |
| Use Fast algorithm instead of Gurobi at size 80 (L2) | Wasserstein DRMDP Bellman update (q=2) · Random DRMDP N=S=A=80 | runtime (s): 44.3305 -> 24.6080 | Improved |
| Use Fast algorithm instead of FOM at size 80 (L2) | Wasserstein DRMDP Bellman update (q=2) · Random DRMDP N=S=A=80 | runtime (s): 31.1326 -> 24.6080 | Improved |
| Use Fast algorithm instead of Gurobi at size 50 (L2) | Wasserstein DRMDP Bellman update (q=2) · Random DRMDP N=S=A=50 | runtime (s): 5.1669 -> 7.3717 | Hurt |
| Use Fast algorithm instead of FOM at size 50 (L2) | Wasserstein DRMDP Bellman update (q=2) · Random DRMDP N=S=A=50 | runtime (s): 9.1936 -> 7.3717 | Improved |
| Use Fast algorithm instead of Gurobi at size 60 (L2) | Wasserstein DRMDP Bellman update (q=2) · Random DRMDP N=S=A=60 | runtime (s): 11.6760 -> 10.8293 | Improved |
Closest papers by meaning
- Efficient Value Iteration for s-rectangular Robust Markov Decision Processes (ICML 2024 · similarity 0.90)
- Online MDP with Prototypes Information: A Robust Adaptive Approach (AAAI 2025 · similarity 0.88)
- Provable Policy Gradient for Robust Average-Reward MDPs Beyond Rectangularity (ICML 2025 · similarity 0.87)
- On the Complexity of Discounted Robust MDPs with $L_p$ Uncertainty Sets (NeurIPS 2026 · similarity 0.87)
- Finite-Sample Analysis of Policy Evaluation for Robust Average Reward Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Soft Robust MDPs and Risk-Sensitive MDPs: Equivalence, Policy Gradient, and Sample Complexity (ICLR 2024 · similarity 0.87)
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.
