基于数学形态学的围棋棋群聚类算法

Cluster Algorithm of Go Groups Based on Mathematical Morphology

计算机围棋可以模拟人类棋手的棋群聚类能力以提高搜索效率。本研究以数学形态学为工具,在形式化基础上采用带有限制条件的膨胀运算进行棋群的初步聚类,结合其它一些启发式搜索方法完成棋群的最终聚类,并结合实战时局评价了此算法的性能,指出了此算法的应用价值。

The computer go can simulate human player＇s go group clustering ability to improve its searching efficiency. This study, taking mathematical morphology as a tool, based on the formalization, adopts constrained dilating operation to undergo preliminary clustering on go groups, and combines with some other heuristic searching methods to complete the final clustering on go groups. This paper, by linking with actual combat, also evaluates the algorithm＇s capability, and at the same time points out its application value.