减少高阶矩阵运算和存储量方法的研究  

The Research of Reducing High Level Matrix Calculation and Storage

作　　者：宋志峰[1] 梅顺齐[1] 

机构地区：[1]武汉科技学院机电工程系,湖北武汉430073

出　　处：《武汉纺织工学院学报》2003年第3期35-38,共4页

摘　　要：高阶矩阵运算和存储量都特别大,为了减少运算和存储量,本文讨论了稀疏矩阵、单位矩阵、对称矩阵的存储方法。介绍采用引用计数、操作标志等方法实现矩阵复制、转置、乘常数等部分矩阵运算。该方法在产生新矩阵时,并不生成数据的拷贝,而是改变矩阵的引用计数、操作标志等变量,生成新的矩阵,减少了运算量和存储量。Calculation and storage are trouble for high level matrix .For the sake of these problems, the storage method of the sparse matrix, unit matrix, symmetry matrix are discussed in this paper. Then the methods of the adoption to quote from to count, operating the marking etc are introduced. If these methods are adopted and data are not copied, but quote from to count and operate the marking etc are changed, calculation and storage are reduced.

关 键 词：高阶矩阵 运算量 存储量 引用计数 操作标志 复制 转置 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]