摘要(Abstract):
为了解决黑盒问题优化领域中传统优化算法在学习问题结构时存在缺乏样本多样性的问题,设计结合近邻传播聚类的世界生成神经网络优化器。该优化器通过独特的世界模型与采样生成器进行协同学习,完成对问题结构的学习并生成更加多样化的解,以此为基础提出新的优化算法;将所提出的算法与5种有代表性的算法在12个不同特征的优化问题上进行多角度的对比。结果表明,结合近邻传播聚类的世界生成神经网络优化器在不同特征基准问题上平均性能达到最优,准确度平均排名第一,证明了结合近邻传播聚类的世界生成神经网络优化器在学习问题结构的有效性,同时增加了样本的多样性。
关键词(KeyWords): 近邻传播聚类;世界生成神经网络;黑盒问题;世界模型;神经网络;协同学习
基金项目(Foundation): 山东省重点研发计划项目(2019GGX101041);; 济南市高校院所科研带头人工作室项目(2021GXRC077)
作者(Author): 陆敏芳,宗伟,陈美涵,杨波,王琳,张波
DOI: 10.13349/j.cnki.jdxbn.20230524.002
参考文献(References):
[1] LIU S J,CHEN P Y,KAILKHURA B,et al.A primer on zeroth-order optimization in signal processing and machine learning:principals,recent advances,and applications[J].IEEE Signal Processing Magazine,2020,37(5):43.
[2] GARG H.A hybrid PSO-GA algorithm for constrained optimization problems[J].Applied Mathematics and Computation,2016,274:292.
[3] 王峰,张衡,韩孟臣,等.基于协同进化的混合变量多目标粒子群优化算法求解无人机协同多任务分配问题[J].计算机学报,2021,44(10):1967.
[4] MARK H,MARTIN P.An introduction and survey of estimation of distribution algorithms[J].Swarm and Evolutionary Computation,2011,1(3):111.
[5] FREY B J,DUECK D.Clustering by passing messages between data points[J].science,2007,315(5814):972.
[6] LI X K,YANG Q Y,WANG Y,et al.Development of surrogate models in reliability-based design optimization:a review[J].Mathematical Biosciences and Engineering,2021,18(5):6386.
[7] 肖汉光,蔡从中.特征向量的归一化比较性研究[J].计算机工程与应用,2009,45(22):117.
[8] HANSEN N,ANNE A,ROS R,et al.COCO:a platform for comparing continuous optimizers in a black-box setting[J].Optimization Methods and Software,2020,36(1):114.
[9] SINAY M,SARAFIAN E,LOUZOUN Y,et al.Explicit gradient learning[EB/OL].(2020-06-09)[2022-04-03].https://doi.org/10.48550/arXiv.2006.08711.
[10] SINGER S,NELDER J.Nelder-mead algorithm[J].Scholarpedia,2009,4(7):2928.
[11] DAI Y H.Convergence properties of the BFGS algorithm[J].SIAM Journal on Optimization,2002,13(3):693.
[12] MARINI F,WALCZAK B.Particle swarm optimization (PSO):a tutorial[J].Chemometrics and Intelligent Laboratory Systems,2015,149:153.
[13] KATOCH S,CHAUHAN S S,KUMAR V.A review on genetic algorithm:past,present,and future[J].Multimedia Tools and Applications,2021,80:8091.