!// 首先給出對稱正定矩陣的概念
!// 如果A'= A,則n*n矩陣A是對稱矩陣。如果對於所有向量x/=0,x'Ax > 0,則稱矩陣A是正定矩陣
!// 對稱正定矩陣是非奇異的
!// 任意矩陣A的行列式爲矩陣對應的特徵值的乘積
!// 對於對稱正定矩陣,可使用楚列斯基分解方法,可分解爲: A = R'R,其中R是一個上三角矩陣
!// 使用楚列斯基分解,對於對稱正定矩陣,和一般的矩陣相比,它們只有一半數量的獨立元素,可以用一半的計算代價實現,並且僅僅使用一半的內存
Module mod
Implicit none
Integer, parameter :: n = 3
Real(kind=8) :: A(n,n) = [ 4.d0, -2.d0, 2.d0, -2.d0, 2.d0, -4.d0, 2.d0, -4.d0, 11.d0 ]
Real(kind=8) :: b(n,1) = reshape([6.d0,-10.d0,27.d0],[n,1])
Real(kind=8) :: R(n,n) = 0.d0, RR(n,n) = 0.d0
Real(kind=8) :: x(n,1) = 0.d0
Contains
Subroutine GetR ()
Implicit none
Integer :: i, j
Real(kind=8), allocatable :: u(:,:)
u = 0.d0
Do i = 1, n
If ( A(i,i) < 0.d0 ) then
Write ( *,'(1x,g0)' ) '矩陣A不是對稱正定矩陣,程序結束!'
stop
End if
R(i,i) = sqrt(A(i,i))
j = n - i
If ( j >= 1 ) then
Allocate( u(j,1) )
u(:,1) = A(i,i+1:n) / R(i,i)
R(i,i+1:n) = u(:,1)
A(i+1:n,i+1:n) = A(i+1:n,i+1:n) - matmul( u,transpose(u) )
Deallocate( u )
End if
End do
RR = transpose(R)
End subroutine GetR
!// Ax = b, A = R'R
!// R'Rx = b, 令R'c = b 求出c
!// 最後用Rx = c求出x
Subroutine GetRoot ()
Implicit none
Integer :: i, j
Real(kind=8) :: c(n,1) = 0.d0
!// R'c = b
Do i = 1, n
Do j = 1, i - 1
b(i,1) = b(i,1) - RR(i,j) * c(j,1)
End do
c(i,1) = b(i,1) / RR(i,i)
End do
!// 求x: Rx = c
Do i = n, 1, -1
Do j = i + 1, n
c(i,1) = c(i,1) - R(i,j) * x(j,1)
End do
x(i,1) = c(i,1) / R(i,i)
End do
Write ( *,'(1x,a)' ) '原方程的解爲:'
Do i = 1, n
Write ( *,'(f12.5)' ) x(i,1)
End do
End subroutine GetRoot
End module
Program CholeskyDecomposition
use mod
Implicit none
call GetR ()
call GetRoot ()
End program CholeskyDecomposition