基于遗传算法的属性约简新方法

属性约简是粗糙集理论的一个核心问题,而求解最小约简是NP-Hard问题。为了有效获取最小相对约简,提出一种基于遗传算法的粗糙集属性约简算法,算法将属性核加入遗传算法的初始种群来增加收敛速度,而且在适应度函数中,引入决策属性对条件属性的依赖度,使算法既保证全局寻优的特性又具有加强局部搜索的能力,能够获得最优的搜索效果。该算法通过实例分析,证明是求解属性约简问题的快速有效方法。

A Novel Attribute Reduction Algorithm Based on GA

Attribute reduction is a key problem for the rough set theory and to solve the minimum reduction belongs to the NP-Hard problem.In order to achieve effectively relative attribute reduction,proposes a rough set attribute reduction algorithm based on GA.The core is joined initial population in GA in order to accelerate capacity.And introduces the dependence degree of decision attribute in the condition attribute in the fitness function,so as the algorithm can not only ensure the global optimization performance but strengthen the local search ability,obtain optimal search performacnce.Experiential result shows the algorithm is fast and effective.