象棋棋形辨識之研究

Abstract

象棋是一個相當受歡迎的棋類遊戲，在世界各地有廣大的玩家，象棋AI程式更是近代一個相當熱門的領域，許多人投入這方面的研究，不論在學術或商業界，都有很多傑出的作品。如果能針對象棋AI程式的設計做探討，使AI能夠提高棋力，對於棋與資訊界都將是很大幫助。

本篇論文研究象棋類AI，把用在西洋棋、將棋與圍棋上的棋形辨識技術，做些探討與改善應用在象棋上面，並提出一套有系統的棋形建置方式，產生出能快速比對棋形的程式碼，最後評估並測量出象棋棋形辨識之效能及其實際應用於象棋AI程式的效果。

Chinese Chess is a very popular game played by many people in the world. And Chinese Chess AI program is researched widely nowadays.

This thesis studies AI Programming for Chinese Chess-like Games and improves the original technology used in Chess, Shogi and Go in order to apply it on Chinese Chess. It also proposes a new systematical process to build patterns. We evaluate and measure the efficiency of Chinese Chess pattern recognition and the actual effect when using it in the Chinese Chess AI program.