Improving Strategic Play in Shogi by Using Move Sequence Trees

نویسنده

  • Reijer Grimbergen
چکیده

The main weakness of shogi programs is considered to be in the opening and the middle game. Deep search is not enough to cover the lack of strategic understanding, so most strong shogi programs use a hill-climbing approach to build castle and assault formations. The problem of a hill-climbing approach is that if the final result of two different paths is the same, then the final score will also be the same. In shogi, this leads to high priority moves being pushed down the principal variation and in extreme cases such moves will never be actually played. To solve this problem, move sequence trees are proposed, which guide the program through the right order of moves. This approach was tested in the yagura opening. A preliminary self-play experiment shows that using move sequence trees could be an enhancement of hill-climbing approaches for strategic play in shogi.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Using Bitboards for Move Generation in Shogi

In this paper it will be explained how to use bitboards for move generation in shogi. In chess, bitboards have been used in most strong programs because of the easy representation of a chess board by a single 64-bit integer. For shogi, a less efficient representation has to be used because a shogi board has 81 squares instead of 64. A representation with an array of three integers is proposed, ...

متن کامل

Pattern Recognition for Candidate Generation in the Game of Shogi

Chess has been an important research area in Arti cial Intelligence for decades. It seems that the strongest programs will soon play better than the best experts, so it is time to look at other game domains as possible test domains for AI research. We feel that shogi is well suited for this, because it is similar to chess, yet signi cantly different. In this paper we give the main differences b...

متن کامل

Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm

The game of chess is the most widely-studied domain in the history of artificial intelligence. The strongest programs are based on a combination of sophisticated search techniques, domain-specific adaptations, and handcrafted evaluation functions that have been refined by human experts over several decades. In contrast, the AlphaGo Zero program recently achieved superhuman performance in the ga...

متن کامل

Enhancing Search Efficiency by Using Move Categorization Based on Various Game Progress Values in Amazons

Amazons is a two-player perfect information game with an average number of legal moves that is higher than chess, shogi or Go. Our aim is to improve the search efficiency of an Amazons program by finding good moves to search first, thus improving the efficiency of α-β search. In earlier work, we established that a static move ordering scheme has important problems and that move ordering needs t...

متن کامل

An Automatic Recording System of the Plays and Moves of SHOGI Games Using Image Processing Technique

This paper describes on a system which records the plays and moves of SHOGI games automatically by using image processing technique from the pictures captured with a digital video camera. We have considered the overall structure of the system, and the recognition of movements of a piece. "simple-move" (ugoku), "promoting" (naru), "dropping-move" (utsu), and "capture-move" (toru). This paper esp...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2007