matlab 用LU分解求解线性方程组——代码记录

转自http://www.cnblogs.com/wangshide/archive/2011/12/09/2282641.html

求解Ax=d的解x

1. LU分解

%  file : myLU.m

function [L,U] = myLU(A) % 实现对矩阵A的LU分解，L为下三角矩阵 A [n,n] = size(A); L = zeros(n,n); U = zeros(n,n); for  i = 1 :n      L(i,i) = 1 ; end for  k = 1 :n      for  j = k:n          U(k,j) = A(k,j) - sum (L(k, 1 :k - 1 ). * U( 1 :k - 1 ,j)');      end      for  i = k + 1 :n          L(i,k) = (A(i,k) - sum (L(i, 1 :k - 1 ). * U( 1 :k - 1 ,k)')) / U(k,k);      end end

2. 用分解获得的L，U求解x

 

[L,U] =  myLU(A)   [n,m] = size(A)   y( 1 ) = d( 1 ); for  i = 2 :n         for  j = 1 :i - 1          d(i) = d(i) - L(i,j) * y(j);      end      y(i) = d(i); end   x(n) = y(n) / U(n,n); for  i = (n - 1 ): - 1 : 1      for  j = n: - 1 :i + 1          y(i) = y(i) - U(i,j) * x(j);      end      x(i) = y(i) / U(i,i); end