Abstract: Attribute reduction is one of the important issues of rough set theory.In order to achieve effectively attribute reduction，a kind of attribute reduction algorithm based on GA is proposed.The conditions attribute is joined relative core according to the importance degree of the condition attribute，and the relative core is joined initial population in GA in order to accelerate convergence.The authors get consequence and decision by using the algorithm to analyze maritime accidences.

Key words: rough set theory, Genetic Algorithms（GA）, fitness function, attribute reduction

摘要： 属性约简是粗糙集合研究的重要内容之一。为了能够有效地获取决策表中属性最小相对约简，提出一种基于遗传算法的属性约简算法。依据条件属性的重要度，把条件属性加入到相对核属性集中，将相对核加入遗传算法的初始种群中以加快算法收敛。并应用该算法对水上事故模型预测进行分析，做出推理和决策。

关键词: 粗糙集, 遗传算法, 适应度函数, 属性约简