在保持一对覆盖上、下近似算子不变的条件下,探讨覆盖族的约简。利用所构造的辩识矩阵给出覆盖族的约简与核心的判别定理,并提出基于信息量的寻找最小约简的算法,从而进一步完善广义覆盖粗集的约简理论。
This paper proposes the reduction for a family of coverings preserving the upper approximation operator and lower approximation operator. Then, it constructs a discernibility matrix to give judgement theorems for reducts and core. It also gives an algorithm based on information quantity for the reduction. The results add something divertivc to covering generalized rough sets.