论文部分内容阅读
遗传算法已经在多目标优化问题中得到了广泛应用及深入研究,NSGA-Ⅱ是求解多目标优化问题的代表算法之一,其中聚集距离在收敛性和分布均匀性上均起到了重要作用,但算法没有充分考虑微观的个体本身和宏观的种群整体的作用.为了能更合理地估计区域密度,使所求解集更好更均匀地收敛于Pareto最优边界,笔者基于均匀聚集区间和基尼权重构造了一种均匀聚集距离算子,并基于该算子提出了一种改进的NSGA-Ⅱ算法.最后,通过对6个标准多目标测试问题的实验验证了算法的有效性.
Genetic algorithms have been widely used and deeply studied in multi-objective optimization problems. NSGA-Ⅱ is one of the representative algorithms for solving multi-objective optimization problems. The clustering distance plays an important role in the convergence and distribution uniformity. However, In order to estimate the regional density more reasonably and make the solution set converge to the Pareto optimal boundary more and more uniformly, the algorithm does not fully consider the microscopic individuals and macro population as a whole. Based on the uniform aggregation interval and the Gini weights, A uniform aggregation distance operator is proposed and an improved NSGA-Ⅱ algorithm is proposed based on the operator.Finally, the effectiveness of the algorithm is verified by experiments on six standard multi-objective test problems.