fortran如何生成高斯矩阵

fortran中生成的一个随机矩阵如何转换成一个均值为0标准差为2的高斯矩阵

http://fcode.cn/code_prof-33-1.html

Module ran_mod
  Implicit None
! ran return a uniform random number between 0-1  
! norma return a normal distribution  
contains 
  function ran()   !returns random number between 0 - 1  
    implicit none 
    integer , save :: flag = 0
    double precision :: ran 
    if(flag==0) then 
      call random_seed()
      flag = 1 
    endif 
    call random_number(ran)     ! built in fortran 90 random number function  
  end function ran

  function normal(mean,sigma) 
    implicit none 
    integer :: flag 
    double precision, parameter :: pi = 3.141592653589793239  
    double precision :: u1, u2, y1, y2, normal, mean, sigma 
    save flag 
    data flag /0/ 
    u1 = ran(); u2 = ran() 
    if (flag.eq.0) then 
      y1 = sqrt(-2.0d0*log(u1))*cos(2.0d0*pi*u2) 
      normal = mean + sigma*y1 
      flag = 1 
    else 
      y2 = sqrt(-2.0d0*log(u1))*sin(2.0d0*pi*u2) 
      normal = mean + sigma*y2 
      flag = 0 
    endif  
  end function normal 
!The above codes are made in Fortran 90 language, if you have any question, you may write to sealin2008@hotmail.com 
End Module ran_mod

Program www_fcode_cn
  use ran_mod
  Implicit None
  Integer , parameter :: N = 10000
  Real( Kind = 8 ) :: a( N )
  integer :: i
  Open( 12 , File = 'out.txt' )
  Do i = 1 , N
    a( i ) = normal( 5.0D0 , 2.0D0 )
    write( 12 , * ) a(i)
  End Do
  Close( 12 )
End Program www_fcode_cn

上述代码可以生成符合高斯分别的数据，矩阵的话，你得做个约束就行

Module ran_mod
Implicit None
! ran return a uniform random number between 0-1

! norma return a normal distribution

contains
function ran() !returns random number between 0 - 1

implicit none
integer , save :: flag = 0
double precision :: ran
if(flag==0) then
call random_seed()
flag = 1
endif
call random_number(ran) ! built in fortran 90 random number function

end function ran

function normal(mean,sigma)
implicit none
integer :: flag
double precision, parameter :: pi = 3.141592653589793239

double precision :: u1, u2, y1, y2, normal, mean, sigma
save flag
data flag /0/
u1 = ran(); u2 = ran()
if (flag.eq.0) then
y1 = sqrt(-2.0d0*log(u1))*cos(2.0d0*pi*u2)
normal = mean + sigma*y1
flag = 1
else
y2 = sqrt(-2.0d0*log(u1))*sin(2.0d0*pi*u2)
normal = mean + sigma*y2
flag = 0
endif

end function normal
!The above codes are made in Fortran 90 language, if you have any question, you may write to sealin2008@hotmail.com
End Module ran_mod

Program www_fcode_cn
use ran_mod
Implicit None
Integer , parameter :: N = 10000
Real( Kind = 8 ) :: a( N )
integer :: i
Open( 12 , File = 'out.txt' )
Do i = 1 , N
a( i ) = normal( 5.0D0 , 2.0D0 )
write( 12 , * ) a(i)
End Do
Close( 12 )
End Program www_fcode_cn

我遇到了同样的问题，请问一下您的问题怎么解决的