摘要
介绍一种新的不动点迭代算法,用于解决矩阵秩最小化问题.此算法是在原不动点算法基础上,将收缩算子与单位算子做一个凸组合,进行加速.并证明了新算法的收敛性.
In this paper, a new fixed point iterative algorithm is introduced, which is to solve the matrix rank minimization problem. Based on the original fixed point iterative algorithm, the new algorithm makes a convex combination between unit operator and shrinkage operator, and accelerates the iterative algorithm successfully. At last, the convergence of the new algorithm is proved.
出处
《哈尔滨师范大学自然科学学报》
CAS
2012年第5期21-23,共3页
Natural Science Journal of Harbin Normal University
关键词
矩阵秩
不动点迭代算法
收缩算子
Matrix rank
Fixed point iterative algorithm
Shrinkage operator
