摘要
针对径向基函数插值方法,提出了一种新的重启动策略来改进径向基的优化效果.在采用SLHD(symmetry Latin Hypercube design,对称拉丁超立方设计)方法换新的初始点对优化效果没有太大改进的时候,提出了一种更换径向基函数的重启动策略,可以取得更好的优化效果.通过数值算例说明了采用这种新的重启动策略在迭代次数上的优越性.
We propose a new restarting strategy to improve the optimization effect of the radial basis func-tion interpolation method.When there is little improvement by using new initial points in the SLHD (sym-metry Latin Hypercube design) method,we propose a restarting strategy of replacing the radial basis func-tion which yields better numerical optimization performance. Then we illustrate the advantages of restar-ting strategy by some numerical examples.
出处
《湖北民族学院学报(自然科学版)》
CAS
2016年第4期398-401,共4页
Journal of Hubei Minzu University(Natural Science Edition)
关键词
径向基函数
全局最优化
响应面模型
重启动策略
对称拉丁超立方设计
radial basis function
global optimization
response surface model
restarting strategy
sym-metry Latin Hypercube design