A study on knowledge-based five-in-a-row

지식 베이스를 이용한 오목게임에 관한 연구

The object of this thesis is to develop an effective method for Five-in-a-Row. For this purpose, we develop a method that describes a move accurately, and a procedure that choses a move more effectively. In describing a move, we represent each move as eight vectors generated from four directions for both offensive and defensive strategies. Each vector consists of three components, that is, connectivity, agressiveness, and directions of stones in a line. In choosing a move, we introduce a preference order among all vectors by an expert's knowledge. Then, an efficient algorithm is developed to test a cycle in the preference order. Experimental results show that the success rate of this new method is about 90% against Weisenbaum's program known as the prototype of the Five-in-a-Row program.