基于种群自适应调整的多目标差分进化算法

发布时间：2018-11-13 11:25

【摘要】：为提高已有多目标进化算法在求解复杂多目标优化问题上的收敛性和解集分布性,提出一种基于种群自适应调整的多目标差分进化算法。该算法设计一个种群扩增策略,它在决策空间生成一些新个体帮助搜索更优的非支配解;设计了一个种群收缩策略,它依据对非支配解集的贡献程度淘汰较差的个体以减少计算负荷,并预留一些空间给新的带有种群多样性的扰动个体;引入精英学习策略,防止算法陷入局部收敛。通过典型的多目标优化函数对算法进行测试验证,结果表明所提算法相对于其他算法具有明显的优势,其性能优越,能够在保证良好收敛性的同时,使获得的Pareto最优解集具有更均匀的分布性和更广的覆盖范围,尤其适合于高维复杂多目标优化问题的求解。
[Abstract]:In order to improve the convergence and solution set distribution of existing multi-objective evolutionary algorithms for solving complex multi-objective optimization problems, a multi-objective differential evolutionary algorithm based on population adaptive adjustment is proposed. The algorithm designs a population expansion strategy, which generates some new individuals in the decision space to help search for better non-dominated solutions. A population shrinking strategy is designed to eliminate individuals with poor contribution to the non-dominated solution set to reduce computational load and to reserve some space for new disturbed individuals with population diversity. The elite learning strategy is introduced to prevent the algorithm from falling into local convergence. The algorithm is tested and verified by typical multi-objective optimization function. The results show that the proposed algorithm has obvious advantages over other algorithms, and its performance is superior, and it can ensure good convergence at the same time. The obtained Pareto optimal solution set has more uniform distribution and wider coverage, which is especially suitable for solving high dimensional complex multi-objective optimization problems.
【作者单位】：东华大学管理学院;
【基金】：国家自然科学基金资助项目(70971020) 上海市自然科学基金资助项目(15ZR1401600)
【分类号】：TP18

【相似文献】