基于改进遗传算法的粗糙集属性约简算法

Attribute reduction algorithm for rough set based on improving genetic algorithm

属性约简是粗糙集理论研究的主要内容之一,为了能够有效地获取决策表中属性最小约简,在分析属性约简的方法与遗传算法的基础上,将属性重要性度量作为启发式信息引入遗传算法,提出了一种启发式遗传算法。通过构造新的变异算子来引入启发式信息,体现了启发式信息的局部搜索技术,使得算法既保持整体优化特性,又具有较快的收敛速度。实验结果表明,该方法能快速有效地求出决策表的最小约简。

An attribute reduction is the main content which the rough set theory studies,and the goal is to achieve the minimal reduciton of the attributes efficitively in a decision table.Based on analysis of attribute reduction and genetic algorithm and regarding the significance of attributes as heuristic information,the heuristic information is introduced into genetic algorithm,and an effective heurisitic genetic algorithm is proposed.A new mutate operator is used for introducing the heurisitic information and the operator is a local research method using heurisitic information.So the algorithm converges quickly and has global optimizing ability.The results show that the method can calculate minimal reduction of decision charts quickly and effectively.