【摘要】 提出一种基于免疫的多目标优化遗传算法。该算法模仿生物免疫系统过程,使用克隆选择算子和高斯变异算子提高了搜索效率和收敛性;创建了一个记忆细胞集来保存每代所产生的Pareto最优解,以便产生Pareto最优解集;提出一种有别于传统聚类算法的邻近排挤算法对记忆细胞集进行不断的更新及删除,保证了Pareto最优解集的分布均匀性。最后将该算法与SPEA算法分别进行了仿真,通过比较两者的收敛性和分布性,得到前者优于后者的结论。更多还原

【Abstract】 The paper presented a kind of Multi-objective Optimization Genetic Algorithm based on Immune(MOGAI).The algorithm used the natural immune systems as reference,in order to improve the searching efficiency and the convergence;Clone operator and Gaussian mutation operator were used in the algorithm;a memory cells set was set up to preserve the Pareto optimal solutions coming from every generation,in order to produce the Pareto optimal solutions set;To assure the distribution of the Pareto optimal solutions set,a vicinity crowding algorithm different from the cluster algorithm was used to update the memory cells set.Finally,simulation was carried on the the algorithm and the SPEA.By comparing the convergence and the distribution of these two algorithms,the paper obtained that MOGIA is better than SPEA.更多还原