Development of Shogi opening algorithm that exceeds the Meijin
Project/Area Number |
24500192
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Intelligent informatics
|
Research Institution | Matsue National College of Technology |
Principal Investigator |
|
Project Period (FY) |
2012-04-01 – 2018-03-31
|
Project Status |
Completed (Fiscal Year 2017)
|
Budget Amount *help |
¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000)
Fiscal Year 2016: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2015: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2014: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2013: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2012: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
|
Keywords | コンピュータ将棋 / 評価項目自動抽出法 / 序盤 / 機械学習 / 将棋 / 評価関数 / Bonanza Method / freq型 / テトリス / Deep Learning / bonanza method / 棋譜 / 名人 |
Outline of Final Research Achievements |
Computer shogi is generally strong in the end game, but there are still weak points in the opening game. Although professional players can evaluate a good "shape", the computer can not evaluate "shape" consisting of many pieces position pattern. In order to realize professional-looking play in the opening game that understands the "shape", evaluation consisting of many pieces positions is required. However, as the number of evaluation factors increases, memory and time increase exponentially, which is difficult to realize. In this research, in order to evaluate a high frequency pattern using multiple frames, we propose "automatic evaluation item evaluation using random numbers" and succeeded in automatically extracting high frequency evaluation factors. However, it was found that considerable speeding up was necessary for practical use, we proposed a new data structure and succeeded in increasing the speed to a certain level.
|
Report
(7 results)
Research Products
(7 results)