In the next section of this paper we describe a non - generational ga for multiobjective optimization problems ( mop ) . in it the replacement policy is such that an offspring replaces the worst one in the current population only if it is better then it 在将错位交叉算子应用于多目标优化问题求解时,本文提出一种基于非群体迭代型的小生境多目标优化遗传算法,即算法迭代过程中不是更换种群全部或大部分个体,而是每次更新群体中最差的个体。