将棋における合議アルゴリズム――多数決による手の選択  [in Japanese] Consultation Algorithm in Shogi --- A Move Decision by Majority  [in Japanese]

Access this Article

Search this Article

Abstract

本論文では,将棋プログラムの新たな並列処理手法を提案する.このアルゴリズムは,複数の思考プログラムの候補手の中から一つの手を選択する手法である.このアルゴリズムを合議アルゴリズムと呼ぶ.本論文では,将棋における合議手法の提案と評価を行い,また単一プログラムからでも乱数を用いた合議手法を提案しその有効性も示した.さらに,YSS,GPS将棋,Bonanza等の有名な強豪プログラムをこの合議アルゴリズムで組み合わせることで,その各々のプログラムよりも強くなることを示した.A new algorithm that runs on a computer with interconnected processors has been designed for Shogi. The algorithm adopts consultation between many individual players. We call the algorithm "Consultation Algorithm". In this paper, we proposed and evaluated the consultation approach in Shogi, and proposed how to generate multiple players from a single program by using the random number and expressed the effectiveness of this method. Applying a simple majority rule to select a move from multiple moves, the consultation algorithm improves the performance of computer Shogi programs. It is also demonstrated that the consultation algorithm consisting of three famous Shogi programs: YSS, GPS and Bonanza plays better games than any of the three programs individually.

A new algorithm that runs on a computer with interconnected processors has been designed for Shogi. The algorithm adopts consultation between many individual players. We call the algorithm "Consultation Algorithm". In this paper, we proposed and evaluated the consultation approach in Shogi, and proposed how to generate multiple players from a single program by using the random number and expressed the effectiveness of this method. Applying a simple majority rule to select a move from multiple moves, the consultation algorithm improves the performance of computer Shogi programs. It is also demonstrated that the consultation algorithm consisting of three famous Shogi programs: YSS, GPS and Bonanza plays better games than any of the three programs individually.

Journal

  • 情報処理学会論文誌

    情報処理学会論文誌 52(11), 3030-3037, 2011-11-15

    情報処理学会

Keywords

Codes

  • NII Article ID (NAID)
    110008677680
  • NII NACSIS-CAT ID (NCID)
    AN00116647
  • Text Lang
    JPN
  • Article Type
    Journal Article
  • ISSN
    1882-7764
  • NDL Article ID
    023333944
  • NDL Call No.
    YH247-743
  • Data Source
    NDL  NII-ELS  IPSJ 
Page Top