Abstract: The minimum reduction of decision tables is important in reduction learning , and it is also a combinatorial optimization problem. Many heuristic searching algorithms are discussed for the difficulty. In effect, however, the efficiency, robustness and completeness of minimum solutions etc. of the algorithms are hard to ensure. Based on rough sets and the extended matrix theory, this paper proposes a solution to overcome the disadvantages using GA(genetic algorithm). Examples show that the method seems to be both effective and can reach the optimal reduction at a high probability.

摘要： 基于粗集理论和扩张矩阵理论，利用遗传算法，提出一种求解最小约简的优化方法。实验说明，此方法能以较大的概率全面获得决策表的最小约简，实用性较强。

关键词: 决策表, 最小约简, 遗传算法, 粗集

CLC Number: 

• TP18