Amplifying Exploration in Monte-Carlo Tree Search by Focusing on the Unknown
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914677699641344 |
|---|---|
| author | Derstroff, Cedric Brugger, Jannis Blüml, Jannis Mezini, Mira Kramer, Stefan Kersting, Kristian |
| author_facet | Derstroff, Cedric Brugger, Jannis Blüml, Jannis Mezini, Mira Kramer, Stefan Kersting, Kristian |
| contents | Monte-Carlo tree search (MCTS) is an effective anytime algorithm with a vast amount of applications. It strategically allocates computational resources to focus on promising segments of the search tree, making it a very attractive search algorithm in large search spaces. However, it often expends its limited resources on reevaluating previously explored regions when they remain the most promising path. Our proposed methodology, denoted as AmEx-MCTS, solves this problem by introducing a novel MCTS formulation. Central to AmEx-MCTS is the decoupling of value updates, visit count updates, and the selected path during the tree search, thereby enabling the exclusion of already explored subtrees or leaves. This segregation preserves the utility of visit counts for both exploration-exploitation balancing and quality metrics within MCTS. The resultant augmentation facilitates in a considerably broader search using identical computational resources, preserving the essential characteristics of MCTS. The expanded coverage not only yields more precise estimations but also proves instrumental in larger and more complex problems. Our empirical evaluation demonstrates the superior performance of AmEx-MCTS, surpassing classical MCTS and related approaches by a substantial margin. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_08511 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Amplifying Exploration in Monte-Carlo Tree Search by Focusing on the Unknown Derstroff, Cedric Brugger, Jannis Blüml, Jannis Mezini, Mira Kramer, Stefan Kersting, Kristian Artificial Intelligence Monte-Carlo tree search (MCTS) is an effective anytime algorithm with a vast amount of applications. It strategically allocates computational resources to focus on promising segments of the search tree, making it a very attractive search algorithm in large search spaces. However, it often expends its limited resources on reevaluating previously explored regions when they remain the most promising path. Our proposed methodology, denoted as AmEx-MCTS, solves this problem by introducing a novel MCTS formulation. Central to AmEx-MCTS is the decoupling of value updates, visit count updates, and the selected path during the tree search, thereby enabling the exclusion of already explored subtrees or leaves. This segregation preserves the utility of visit counts for both exploration-exploitation balancing and quality metrics within MCTS. The resultant augmentation facilitates in a considerably broader search using identical computational resources, preserving the essential characteristics of MCTS. The expanded coverage not only yields more precise estimations but also proves instrumental in larger and more complex problems. Our empirical evaluation demonstrates the superior performance of AmEx-MCTS, surpassing classical MCTS and related approaches by a substantial margin. |
| title | Amplifying Exploration in Monte-Carlo Tree Search by Focusing on the Unknown |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2402.08511 |