以基因演算法進行星海爭霸之單位行動選擇最佳化

Abstract

電腦遊戲從未在電子資訊科技產業中缺席，在這麼多遊戲中即時戰略遊戲一直都是電競比賽的熱門焦點，而星海爭霸可以視為即時戰略遊戲的代表作，可惜目前市面上的即時戰略遊戲，大多利用特殊方法造成玩家與電腦之間條件不對等，達到難易度的調整。所以我們提出了一個構想，使得星海爭霸的遊戲人工智慧變得更強，讓即時戰略遊戲變得更好玩。以外在論文的最後列出了幾個未來可以供人使用的研究方向，希望此研究成為更強的星海爭霸遊戲人工智慧的基石。

Computer game have never been absent in Information technology industry. In so many type of game, the real-time strategy(RTS) game has always been the focus of gaming competitions and StarCraft can be regarded as a masterpiece of real-time strategy game. Unfortunately, most of the real-time strategy game AI can't achieve the same level of player. In order to make RTS game better, we attempt to develop a mechanism which is able to optimize action selection in StarCraft. Because we hope this study become the cornerstone of a great StarCraft AI, we listed some future research directions at the end of this paper.