...
首页> 外文期刊>Journal of Computers >A Trust Region Algorithm Using Curve-Linear Searching Direction for Unconstrained Optimization
【24h】

A Trust Region Algorithm Using Curve-Linear Searching Direction for Unconstrained Optimization

机译:一种使用曲线线性搜索方向的信任区域算法,用于无约束优化

获取原文
   

获取外文期刊封面封底 >>

       

摘要

—In the paper, aimed at the shortcoming of trustregion method, we proposed a algorithm using negativecurvature direction as its searching direction. Theconvergence of the algorithm was given. Furthermore,combing trust region method and curve-linear searchingtechniques, a trust region algorithm, using general curvelinearsearching direction, was proposed. We proved itsefficiency and feasibility. The algorithm has adjustabilityand can select or update its searching direction according tothe iteration. This allows the algorithm that has theproperties of curve-linear searching method and the globalconvergence of trust region method. Finally, we indicatethat some searching directions of common methods can beas a special searching direction of the general method.
机译:- 在纸质中,旨在避风方法的缺点,我们提出了一种使用否定的方向作为搜索方向的算法。给出了算法的变率。此外,提出了梳理信任区域方法和曲线 - 线性搜索技术,使用一般曲线图书化方向的信任区域算法。我们证明了ITSPE,可行性。该算法具有可调节性和可根据迭代选择或更新其搜索方向。这允许该算法具有曲线线性搜索方法的特性和信任区域方法的GlobalConvergentive。最后,我们指示一些搜索方向的常用方法可以让常规方法的特殊搜索方向。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号