Research map / Bandits and online learning / RL theory and MDPs
Minimax Regret Optimisation for Robust Planning in Uncertain Markov Decision Processes
Marc Rigter, Bruno Lacerda, Nick Hawes
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 the regret Bellman recursion rather than myopic CEMR | Minimax regret planning · Disaster rescue, medical decision making, and underwater glider | maximum regret | Improved |
| Use larger n-step options to retain uncertainty dependence over longer action sequences (ablation) | reg · Disaster rescue, medical decision making, and underwater glider | maximum regret | Improved |
| Use stochastic rather than deterministic option policies (ablation) | reg · Disaster rescue and underwater glider | maximum regret | Improved |
| Use stochastic rather than deterministic option policies (ablation) | reg · Disaster rescue and underwater glider | computation time | Hurt |
| Extend CEMR to use n-step options (ablation) | CEMR · Dependent-uncertainty domains | maximum regret | Improved |
| Use action pruning (ablation) | reg, CEMR, and MILP · Evaluation domains | computation time | Improved |
| Use action pruning (ablation) | reg, CEMR, and MILP · Evaluation domains | performance | No clear effect |
| Use minimax regret with n-step options rather than the strongest reported baseline | reg · Disaster rescue and medical decision making | maximum regret on the larger test set | Improved |
| Use minimax regret with larger n-step options | reg · Underwater glider | maximum regret on the larger test set | Improved |
| Use the proposed approach rather than MILP | Minimax regret planning · Medical decision making | solution availability within the time limit: 600s | Improved |
| Use larger n-step options rather than MILP | reg · Dependent-uncertainty domains | maximum regret | Improved |
| Claim exact minimax regret for independent uncertainty sets using the n=1 adversary formulation | Proposed algorithm · Independent uncertainty sets | exactness | Hurt |
Closest papers by meaning
- A Single-Loop Robust Policy Gradient Method for Robust Markov Decision Processes (ICML 2024 · similarity 0.89)
- On the Complexity of Discounted Robust MDPs with $L_p$ Uncertainty Sets (NeurIPS 2026 · similarity 0.89)
- Solving Robust Markov Decision Processes: Generic, Reliable, Efficient (AAAI 2025 · similarity 0.89)
- Efficient Value Iteration for s-rectangular Robust Markov Decision Processes (ICML 2024 · similarity 0.89)
- Achieving Tractable Minimax Optimal Regret in Average Reward MDPs (NeurIPS 2024 · similarity 0.88)
- Provable Policy Gradient for Robust Average-Reward MDPs Beyond Rectangularity (ICML 2025 · similarity 0.88)
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.
