Research map / Learning theory and algorithms / Combinatorial and search optimization
Extreme Value Monte Carlo Tree Search for Classical Planning
Masataro Asai, Stephen Wissow
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 Gaussian bandit (UCB1-Normal2) with UCB1-Uniform | GUCT · IPC subset, hFF, 10^4 node evals | instances solved: 582.95 -> 606.4 | Improved |
| Replace UCB1-Normal2 with UCB1-Uniform | GUCT · IPC subset, hmax | instances solved: 316.6 -> 455.6 | Improved |
| Use Full Bellman backup instead of Monte Carlo backup (ablation) | GUCT with UCB1-Normal2 · IPC subset, hFF | instances solved: 582.95 -> 567.2 | Hurt |
| Use Full Bellman backup instead of Monte Carlo backup (ablation) | GUCT with UCB1-Normal2 · IPC subset, hmax | instances solved: 316.6 -> 263 | Hurt |
| Use Full Bellman backup instead of Monte Carlo backup (ablation) | GUCT with UCB1 (base) or UCB1-Normal · IPC subset, hFF | instances solved: 412 -> 459.4; 283.4 -> 318.8 | Improved |
| Replace GBFS with GUCT-Uniform | GBFS · IPC subset, hFF | instances solved: 538 -> 606.4 | Improved |
| Replace Softmin-Type(h) with GUCT-Uniform | Softmin-Type(h) · IPC subset, hFF | instances solved: 576 -> 606.4 | Improved |
| Replace CHK-Uniform (asymptotically optimal) with UCB1-Uniform | GUCT · IPC subset, hFF | instances solved: 375.4 -> 606.4 | Improved |
| Use Max-k bandits instead of UCB1-Uniform | GUCT · IPC subset, hFF | instances solved: 606.4 -> 162.4 to 267.8 | Hurt |
| Use max-of-Gaussians (+) backup instead of Monte Carlo backup (ablation) | GUCT-Normal2 · IPC subset | instances solved: hFF 582.95 -> 578; hmax 316.6 -> 442.4 | Mixed |
| Add preferred operators (ablation) | GUCT-Uniform · IPC subset, hFF | instances solved: 606.4 -> 635.6 | Improved |
| Use GUCT-Uniform instead of GUCT-Normal2 in Fast Downward | GUCT · IPC 2018 | instances solved: 81.6 -> 80.0 | No clear effect |
Closest papers by meaning
- Online Robust Reinforcement Learning Through Monte-Carlo Planning (ICML 2025 · similarity 0.88)
- Threshold UCT: Cost-Constrained Monte Carlo Tree Search with Pareto Curves (AAAI 2025 · similarity 0.87)
- Monte Carlo Tree Search in the Presence of Transition Uncertainty (AAAI 2024 · similarity 0.87)
- MALinZero: Efficient Low-Dimensional Search for Mastering Complex Multi-Agent Planning (NeurIPS 2025 · similarity 0.86)
- Power Mean Estimation in Stochastic Continuous Monte-Carlo Tree Search (ICML 2025 · similarity 0.86)
- Exploring Starts Are Not Enough: Counterexamples and a Fix for Monte Carlo Exploring Starts (NeurIPS 2026 · 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.
