摘要
迟滞系统广泛存在于各工程领域,但由于迟滞非线性系统的不确定性、状态不可测等特性,因此迟滞系统在建模方面存在一定的困难。针对上述问题,提出了一种采用最小二乘支持向量回归机的解决方案,对系统进行建模方法的研究,并利用粒子群算法、量子粒子群算法等对最小二乘支持向量机中的惩罚参数γ和核函数参数σ的组合进行优化,以提高模型性能及泛化能力。仿真结果表明,利用粒子群优化算法的最小二乘支持向量回归机对迟滞系统的模型仿真可以得到较好的结果。
Hysteresis system widely exists in various engineering fields, but it has certain difficulty in modeling because of uncertainty and undetectable of the hysteresis nonlinear system. In view of the above questions, we proposed a method of least squares support vector regression machine (LS-SVR) to solve the difficulty of nonlinear systems in modeling, and optimizing the parameter γ and σ in LS-SVR using some optimization algorithm, such as particle swarm optimization(PSO) and quantum particle swarm optimization (QPSO) , to improve performance and generalization ability of models. The experimental results show that the hysteresis system modeling can get good results based on QPSO-LSSVR.
出处
《计算机仿真》
CSCD
北大核心
2015年第3期398-402,共5页
Computer Simulation
基金
山东省优秀中青年科学家科研奖励基金(BS2012DX007)
上海市博士后科研资助计划项目(12R21414300)
关键词
迟滞非线性系统
建模
最小二乘支持向量回归机
粒子群算法
量子粒子群算法
Hysteresis nonlinear system
Modeling
Least squares-support vector regression machine (LS-SVR)
Particle swarm optimization(PSO)
Quantum particle swarm optimization(QPS0)