Numerical methods for solving some matrix feasibility problems

详细信息    查看全文

• 作者：Xue-Feng Duan ; Chun-Mei Li ; Jiao-Fen Li ; Yong Ding
• 关键词：Matrix feasibility problem ; Quantum information science ; Projection formula ; Numerical method
• 刊名：Numerical Algorithms
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：74
• 期：2
• 页码：461-479
• 全文大小：
• 刊物类别：Computer Science
• 刊物主题：Numeric Computing; Algorithms; Algebra; Theory of Computation; Numerical Analysis;
• 出版者：Springer US
• ISSN：1572-9265
• 卷排序：74

文摘

In this paper, we design two numerical methods for solving some matrix feasibility problems, which arise in the quantum information science. By making use of the structured properties of linear constraints and the minimization theorem of symmetric matrix on manifold, the projection formulas of a matrix onto the feasible sets are given, and then the relaxed alternating projection algorithm and alternating projection algorithm on manifolds are designed to solve these problems. Numerical examples show that the new methods are feasible and effective.