基于新的初始化策略计算MVEE的积极集算法

发布时间：2018-06-29 00:19

  本文选题：最小体积闭包椭球 + 初始化策略　； 参考：《吉林大学学报(理学版)》2017年05期

【摘要】：针对计算最小体积闭包椭球(MVEE)的积极集算法中原初始化策略耗时较多的问题,先给出一个基于样本协方差矩阵构造的新初始化策略,然后将该初始化策略应用于秩-2更新算法中,并给出一个计算MVEE改进的积极集算法.数值实验结果表明,基于新的初始化策略的积极集算法能有效提高求解大规模数据集MVEE问题的计算效率.
[Abstract]:In order to solve the problem that the original initialization strategy is time-consuming in the active set algorithm for calculating minimum volume closure ellipsoid (MVEE), a new initialization strategy based on sample covariance matrix is proposed. Then, the initialization strategy is applied to the rank -2 update algorithm, and an improved active set algorithm for calculating MVEE is presented. Numerical results show that the active set algorithm based on the new initialization strategy can effectively improve the computational efficiency of solving the MVEE problem of large data sets.
【作者单位】： 西安邮电大学理学院;
【基金】：国家自然科学基金(批准号:11601420;11301415) 陕西省教育厅专项科研计划项目(批准号:15JK1651)
【分类号】：O224
，

本文编号：2079993