摘要
针对社会学习粒子群算法存在的收敛速度慢及进化后期种群多样性缺失等问题,提出了一种基于分层学习的改进粒子群算法。首先,引入分层学习策略,并将其加入社会学习粒子群算法中,实现对种群中不同状态粒子的差别对待,从而增强算法中粒子的探索与开发能力;其次,对个体设定贡献值度量,在贡献值的基础通过减少种群数量,减少计算资源的浪费。最后,使用CEC2010测试函数集对所提算法进行测试,并与5种典型算法进行对比,验证了所提算法的有效性。
In order to solve the problems of social learning particle swarm optimization,such as slow convergence speed and loss of population diversity in late evolution,a particle swarm optimization algorithm based on hierarchical learning is proposed.Firstly,the hierarchical learning strategy is introduced and added to the social learning particle swarm optimization algorithm to realize the differential treatment of different state particles in the population,so as to enhance the exploration and exploitation ability of particles.Secondly,the contribution strategy was designed to avoid the waste of computing resources by reducing the population number.Finally,the algorithm was tested on the CEC2010 benchmark function.The effectiveness of the proposed algorithm is demonstrated by comparing with five algorithms.
作者
白晓慧
何小娟
孙超利
张国晨
BAI Xiao-hui;HE Xiao-juan;SUN Chao-li;ZHANG Guo-chen(Taiyuan University of Science and Technology,Taiyuan 030024,China)
出处
《太原科技大学学报》
2021年第3期169-174,共6页
Journal of Taiyuan University of Science and Technology
基金
国家自然科学基金(61876123)
山西省自然科学基金(201801D121131)
山西省优秀人才科技创新项目(201805D 211028)
山西省留学回国人员科技活动择优资助项目
太原科技大学校博士启动基金(20162029)。
关键词
大规模优化问题
粒子群算法
分层学习策略
贡献值策略
large-scale optimization problem
particle swarm optimization
hierarchical learning strategy
contribution strategy
