Mathematical Sciences: Block, Parallel and Nested Iterative Methods

数学科学：块、并行和嵌套迭代方法

• 批准号: 8807338
• 负责人: Daniel Szyld
• 金额: $ 6.04万
• 依托单位: Duke University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1988
• 资助国家: 美国
• 起止时间: 1988-07-01 至 1990-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8807338&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Block; Parallel; Nested

This research will study convergence properties and implementation details for several types of iterative methods for solving systems of linear equations on sequential computers and on the BBN Butterfly parallel computer. It will examine ways to dynamically determine the number of inner iterations that should be taken in a "nested iterative method," and provide convergence conditions for such methods. Block chaotic algoritms will also be investigated. The question of distributing the computational load in a parallel computing environment is crucial. The proposed research should give insight on this issue.

本研究将研究收敛特性， 几种迭代方法的实现细节， 在顺序计算机上求解线性方程组， 在BBN蝴蝶并行计算机上。 它将研究如何 动态地确定内部迭代的数量， 采用“嵌套迭代法”，并提供收敛性 这些方法的条件。 块混沌算法也将 追究 将计算负荷分布在一个 并行计算环境至关重要。 拟议 研究应使人们了解这一问题。