Research map / Reinforcement learning / MDPs and planning
Deep Reactive Policies for Planning in Stochastic Nonlinear Domains
Thiago P. Bueno, Leliane N. de Barros, Denis D. Mauá, Scott Sanner
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 offline-trained DRP instead of online tf-plan | tf-mdp · Nav2 | inference time per step (s): 1.2 ± 0.2 -> 2.4 · 10−5 | Improved |
| use offline-trained DRP instead of online tf-plan | tf-mdp · Res30 | inference time per step (s): 7.4 ± 1.5 -> 5.5 · 10−5 | Improved |
| deep narrow DRP (tf-mdp 2) instead of shallow wide (tf-mdp 1) (ablation) | tf-mdp · Navigation and HVAC | average total cost | Improved |
| shallow wide DRP (tf-mdp 1) instead of deep narrow (ablation) | tf-mdp · Reservoir | average total cost | Improved |
| DRP vs online tf-plan on solution quality | tf-mdp · Nav2, HVAC3, Res20 | average total cost | No clear effect |
| deep narrow DRP instead of shallow wide on total computing time (ablation) | tf-mdp · HVAC3 | total computing time (s): 138.3 ± 2.2 -> 93.0 ± 2.5 | Improved |
| DRP offline total time vs tf-plan (25) on large reservoir | tf-mdp · Res30 | total computing time (s): 587.7 ± 5.8 -> 344.9 ± 3.1 | Improved |
| DRP offline total time vs tf-plan (10) on Reservoir | tf-mdp · Res30 | total computing time (s): 296.3 ± 4.5 -> 344.9 ± 3.1 | Hurt |
| apply layer normalization to all layers (ablation) | tf-mdp · domains | training benefit | No clear effect |
Closest papers by meaning
- Sample-Efficient Iterative Lower Bound Optimization of Deep Reactive Policies for Planning in Continuous MDPs (AAAI 2022 · similarity 0.92)
- Diffusion Policy Policy Optimization (ICLR 2025 · similarity 0.87)
- Learning Optimal Deterministic Policies with Stochastic Policy Gradients (ICML 2024 · similarity 0.87)
- The Ladder in Chaos: Improving Policy Learning by Harnessing the Parameter Evolving Path in A Low-dimensional Space (NeurIPS 2024 · similarity 0.86)
- Deep learning for continuous-time stochastic control with jumps (NeurIPS 2025 · similarity 0.86)
- Learning Logic Specifications for Policy Guidance in POMDPs: an Inductive Logic Programming Approach (AAAI 2025 · similarity 0.86)
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.
