基于免疫的粗糙集属性约简算法及应用

Immune-based rough sets attribute reduction algorithm and its application

地下施工中影响施工的风险影响因素十分繁杂,为了从中剔除不必要或不重要的因素,提出一种新的基于免疫的粗糙集属性约简算法——IRSAR。该算法在求出决策表的相对核core的基础上初始化抗体群,并定义了新的亲和度计算函数和克隆增殖函数,有效地提高了亲和度的收敛速度。利用记忆数组存放每一代中满足条件的抗体,制定了记忆数组更新策略,从而得到最优抗体。实验结果表明,IRSAR算法能够较快地得出合理有效的约简结果。

The risk factors of underground engineering construction are very complex. In order to remove unnecessary factors, a new immune-based rough sets attribute reduction algorithm （IRSAR） is presented. After finding out the relative core of decision-making table, algorithm initialized the antibodies. A new affinity function and a clone expansion function that effectively improved the convergence speed of affinity are defined. The memory array is used to save those qualified antibodies in each period. An update strategy of memory array is produced to find out the optimal antibody. The experimental results show that IRSAR could get the effective reduction results quickly.