项目名称: 结构矩阵计算的扰动理论与算法研究
项目编号: No.11471279
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 数理科学和化学
项目作者: 黄荣
作者单位: 湘潭大学
项目金额: 62万元
中文摘要: 结构矩阵计算的扰动理论与算法研究具有十分重要的理论意义与应用价值,但存在很多挑战性问题。本项目将致力于探讨结构矩阵计算问题,重点研究矩阵结构分析、结构扰动分析、结构算法设计与分析等论题。我们将通过矩阵结构分析,研究矩阵模式结构和代数结构的参数化、非线性结构的线性化以及结构的保持性等问题;进而以此为基础,研究矩阵计算问题的参数结构扰动分析与非线性结构扰动分析,获得一些强扰动理论成果,并探讨结构扰动与非结构扰动之间的相同与不同;由此,设计与分析相应的结构算法,给出正如结构扰动所确定的强稳定的、高精确的数值计算。我们的目标是解决目前该研究领域存在的一些问题,获得一些有价值的研究成果,发展结构矩阵计算的思想与方法。
中文关键词: 结构矩阵;矩阵计算;扰动理论;数值算法;特征值
英文摘要: The study of perturbation theory and algorithms of structured matrix computations plays an important role in many applications, but there are many challenging issues. The project will work hard to explore structured matrix computation problems by focusing on topics such as matrix structure analysis, structured perturbation analysis, structured algorithms and other topics. We perform matrix structure analysis to parameterize and linearize structured matrices, and we also present structured-preserving properties in the course of numerical computations. We then investigate parameter structured perturbation analysis and nonlinear structured perturbation analysis of matrix computations to obtain some strong perturbation results. In addition, we explore the identity and difference between structured and unstructured perturbations. Finally, we design the corresponding structured algorithms to achieve strong stable and high accurate numerical computations, as shown by structured perturbation results. Our aim is to settle some problems in this field, obtain some valuable research results, and develop ideas and methods of structured matrix computations.
英文关键词: Structured matrix;matrix computation;perturbation theory;numerical algorithm;eigenvalue