象棋棋規基於棋理與邏輯推演法則來解決對弈時出現的循環盤面問題，屬於科學學術研究領域。但是對棋規的理論研究與論著卻罕如鳳毛麟角。近期隨著電腦與網路的盛行，網路象棋對弈平台林立，上網對弈蔚為風潮，卻因棋規複雜而缺乏正確、完整的線上機器裁判機制，導致影響棋局出現循環盤面時機器判決的正確性。 我們經過多年的研究與實驗，本著求真求善的精神，首先對現有的幾種流行棋規進行比較，歸納異同，選擇亞洲棋規為主要研究標的。然後以邏輯檢驗亞洲棋規糾正前後矛盾的規則描述並證明五個棋圖範例的判決錯誤，提供亞洲象棋聯合會未來修訂棋規之學術文獻參考。再根據邏輯歸納法則簡併棋規條文，在確保裁判結果完全正確的要求下，成功的將六十多條術語解釋、棋例總剛及細則簡化為十幾條的條文。最後將規則及裁判流程化，並在個人電腦實作一具有即時裁決能力之電腦象棋裁判系統──「弈仲」來檢驗本文之論述。其主要功能在裁決出現循環盤面之勝、負、和，以螢幕及滑鼠為輸出/入介面。經實驗，「弈仲」能即時且正確的裁判循環盤面，證明本文簡化棋規及棋規程式化的研究結果正確；並可供作網路象棋對弈平台機器仲裁之邏輯核心，提高網路對弈品質。 It is interesting to play Chinese chess games on internet, but the Chinese chess rules are very complex so there is no computer judgment system with complete rules on these Chinese chess game platform. All about Chinese chess theory and logical deduction on Chinese chess rules results from “loop deadlock” problems which belongs to scientific research field. It is hardly to see Chinese chess research and papers about it. Nowadays, the games on computers and internet becomes popular, the platform of internet Chinese chess spread existed. The problem is lacking of online judge by correcting and integrity on complexity Chinese chess rules. The assay is basic for the necessity for judging the “loop deadlock” on line Chinese chess game. In this paper, we examine and correct the contradiction of Asia Chinese chess rules and then prove five forms correctly. I describe the whole picture my rule. It provide to Asia Xiangi Federation (AXF) for revising rules in the future. The success on briefing twelve rules from sixty-four rules after consolidation according to the logical generalization. According to above mention methods make the judge rules online and name it “Yizhong”. The main function of my theory is to judge the “loop deadlock” to be “win”, “fail” or “draw”. The “Yizhong” makes the “loop deadlock” correctly and promptly. Above mentioned content can prove the Chinese chess rules and programming correctly. The aim of the assay provide logical theory online and quality of online Chinese chess game. This system works under AMD 1800 XP CPU and 512 MB RAM. The performance of system can judge in 0.3 second after repetition of position. The judgment is absolutely correct.
