基于稀松密度的多目标优化演化算法

该算法定义和使用稀松密度来保持群体中个体的均匀分布,并将个体的Pareto强度和稀松密度合并到个体的适应值定义中,使得搜索向Pareto最优解集的方向进行并防止早熟;算法还采用多父体杂交策略,每代只产生一个新个体、淘汰一个最差个体,精英个体自动留存。算法复杂度小,占用存储空间少。通过对测试函数的实验,验证了算法的可行性和有效性。

A Multi-objective Evolutionary Algorithm Based on Loosing-Density

To maintain a good spread of solution in the population, the Loosing- density is defined and used in this algorithm,the fitness of the individual through Pareto strength and Loosing- density is also defined . And the strategy of multi- father hybrid is used for producing a new individual,eliminating a old individual and saving the elite individual in each generation .The algorithms has not only less arithmetic but used less memory space. Test results show that the new approach is feasible and effective.