• xmlui.mirage2.page-structure.header.title
    • français
    • English
  • Help
  • Login
  • Language 
    • Français
    • English
View Item 
  •   BIRD Home
  • LAMSADE (UMR CNRS 7243)
  • LAMSADE : Publications
  • View Item
  •   BIRD Home
  • LAMSADE (UMR CNRS 7243)
  • LAMSADE : Publications
  • View Item
JavaScript is disabled for your browser. Some features of this site may not work without it.

Browse

BIRDResearch centres & CollectionsBy Issue DateAuthorsTitlesTypeThis CollectionBy Issue DateAuthorsTitlesType

My Account

LoginRegister

Statistics

Most Popular ItemsStatistics by CountryMost Popular Authors
Thumbnail

Enhancing Playout Policy Adaptation for General Game Playing

Sironi, Chiara; Cazenave, Tristan; Winands, Mark (2021), Enhancing Playout Policy Adaptation for General Game Playing, in Cazenave, Tristan; Teytaud, Olivier; Winands, Mark H. M., Monte Carlo Search, Springer, p. 116-139. 10.1007/978-3-030-89453-5_9

View/Open
EnhancingPPAForGGP.pdf (407.6Kb)
Type
Communication / Conférence
Date
2021
Conference title
First Workshop, MCS 2020, Held in Conjunction with IJCAI 2020
Conference date
2021-01
Conference city
virtuel
Book title
Monte Carlo Search
Book author
Cazenave, Tristan; Teytaud, Olivier; Winands, Mark H. M.
Publisher
Springer
ISBN
978-3-030-89453-5
Number of pages
141
Pages
116-139
Publication identifier
10.1007/978-3-030-89453-5_9
Metadata
Show full item record
Author(s)
Sironi, Chiara
Cazenave, Tristan
Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision [LAMSADE]
Winands, Mark
Abstract (EN)
Playout Policy Adaptation (PPA) is a state-of-the-art strategy that has been proposed to control the playouts in Monte-Carlo Tree Search (MCTS). PPA has been successfully applied to many two-player, sequential-move games. This paper further evaluates this strategy in General Game Playing (GGP) by first reformulating it for simultaneous-move games. Next, it presents five enhancements for the strategy, four of which have been previously successfully applied to a related MCTS playout strategy, the Move-Average Sampling Technique (MAST). Experiments on a heterogeneous set of games show three enhancements to have a positive effect on PPA: (i) updating the policy for all players proportionally to their payoffs instead of updating only the policy of the winner, (ii) collecting statistics for N-grams of moves instead of single moves only, and (iii) discounting the backpropagated payoffs depending on the depth of the playout. Results also show enhanced PPA variants to be competitive with MAST for small search budgets, and better for larger search budgets. The use of an ϵ -greedy selection of moves and of after-move decay of statistics, instead, seem to have a detrimental effect on PPA.
Subjects / Keywords
Monte-Carlo Tree Search; Playout policy adaptation; General game playing

Related items

Showing items related by title and author.

  • Thumbnail
    Combining UCT and Nested Monte-Carlo Search for Single-Player General Game Playing 
    Méhat, Jean; Cazenave, Tristan (2010) Article accepté pour publication ou publié
  • Thumbnail
    Automated Generation of New Concepts from General Game Playing 
    Sato, Yuichiro; Cazenave, Tristan (2014) Communication / Conférence
  • Thumbnail
    Ary, a general game playing program 
    Cazenave, Tristan; Méhat, Jean (2010) Communication / Conférence
  • Thumbnail
    Distributed Nested Rollout Policy for Same Game 
    Negrevergne, Benjamin; Cazenave, Tristan (2017) Communication / Conférence
  • Thumbnail
    A General Multi-Agent Modal Logic K Framework for Game Tree Search 
    Saffidine, Abdallah; Cazenave, Tristan (2012) Communication / Conférence
Dauphine PSL Bibliothèque logo
Place du Maréchal de Lattre de Tassigny 75775 Paris Cedex 16
Phone: 01 44 05 40 94
Contact
Dauphine PSL logoEQUIS logoCreative Commons logo