The Dots and Boxes Records Storing Standard Format for Machine Learning and The Design and Implementation of Its Generation Tool. Issue 3 (March 2019)

• Record Type: Journal Article
• Authors: Gao, Ming; Li, Shuqin; Ding, Meng; Kun, Meng
• Is Part Of: Journal of physics. Volume 1176:Issue 3(2019)
• Publication Date: 2019-03
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1176/3/032009
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: Dots and Boxes is a very popular game for two players. There are many teams in the world who are studying smart game programs about this game. Unfortunately, different from the popular game Go, there are no uniform Dots and Boxes game storing records standards format. This results in a large number of classical match records that cannot be preserved, and it also affects the study of this game algorithms. In particular, machine learning algorithms rely on a large number of records. The AlphaGo developed by Google defeated human top players for the first time in Go with the help of a large number of human chess games. In order to facilitate the recording and saving of Dots and Boxes game records. Based on the JSON standard, this paper designs a Dots and Boxes game records standard format and its storage standard, which has excellent expansion and strong compatibility. It also supports cross-platform applications, facilitates secondary development, and is easy to read and write. In order to facilitate the use and promotion of this standard format, simplify the development process of the game AI algorithm. This article uses Python3 language, based on PyQt5 framework, to implement a Dots and Boxes game software DotsAndBoxes, which can generate and read the above-mentioned game records standard file. It has the characteristics of strong versatility, high portability, good scalability, and low maintenance cost.