Eigenvalues and eigenvectors of a real symmetric matrix in Fortran

In summary: The program tries to solve the system of equations a*x = lambda*x. The system has three equations. The first
  • #1
alui
4
0

Homework Statement



I try to run this program, but there are still some errors, please help me to solve this problems

Homework Equations

The Attempt at a Solution


Code:
        Program Main
!====================================================================
!  eigenvalues and eigenvectors of a real symmetric matrix
!  Method: calls Jacobi
!====================================================================
        implicit none
        integer, parameter :: n=3
        double precision :: a(n,n), x(n,n)
        double precision, parameter:: abserr=1.0e-09
        integer i, j

! matrix A
         data (a(1,i), i=1,3) /   1.0,  2.0,  3.0 /
         data (a(2,i), i=1,3) /   2.0,  2.0, -2.0 /
         data (a(3,i), i=1,3) /   3.0, -2.0,  4.0 /

! print a header and the original matrix
        write (*,200)
        do i=1,n
        write (*,201) (a(i,j),j=1,n)
        end do

        call Jacobi(a,x,abserr,n)

! print solutions
        write (*,202)
        write (*,201) (a(i,i),i=1,n)
        write (*,203)
        do i = 1,n
        write (*,201)  (x(i,j),j=1,n)
        end do

200     format (' Eigenvalues and eigenvectors (Jacobi method) ',/, &
            ' Matrix A')
201         format (6f12.6)
202         format (/,' Eigenvalues')
203         format (/,' Eigenvectors')
            end program main

            subroutine Jacobi(a,x,abserr,n)
!===========================================================
! Evaluate eigenvalues and eigenvectors
! of a real symmetric matrix a(n,n): a*x = lambda*x
! method: Jacoby method for symmetric matrices
! Alex G. (December 2009)
!-----------------------------------------------------------
! input ...
! a(n,n) - array of coefficients for matrix A
! n      - number of equations
! abserr - abs tolerance [sum of (off-diagonal elements)^2]
! output ...
! a(i,i) - eigenvalues
! x(i,j) - eigenvectors
! comments ...
!===========================================================
        implicit none
        integer i, j, k, n
        double precision a(n,n),x(n,n)
        double precision abserr, b2, bar
        double precision beta, coeff, c, s, cs, sc

! initialize x(i,j)=0, x(i,i)=1
! *** the array operation x=0.0 is specific for Fortran 90/95
          x = 0.0
          do i=1,n
          x(i,i) = 1.0
          end do

! find the sum of all off-diagonal elements (squared)
           b2 = 0.0
           do i=1,n
           do j=1,n
           if (i.ne.j) b2 = b2 + a(i,j)**2
           end do
           end do

           if (b2 <= abserr) return

! average for off-diagonal elements /2
          bar = 0.5*b2/float(n*n)

          do while (b2.gt.abserr)
          do i=1,n-1
          do j=i+1,n
          if (a(j,i)**2 <= bar) cycle  ! do not touch small elements
          b2 = b2 - 2.0*a(j,i)**2
          bar = 0.5*b2/float(n*n)
! calculate coefficient c and s for Givens matrix
      beta = (a(j,j)-a(i,i))/(2.0*a(j,i))
      coeff = 0.5*beta/sqrt(1.0+beta**2)
      s = sqrt(max(0.5+coeff,0.0))
      c = sqrt(max(0.5-coeff,0.0))
! recalculate rows i and j
      do k=1,n
        cs =  c*a(i,k)+s*a(j,k)
        sc = -s*a(i,k)+c*a(j,k)
        a(i,k) = cs
        a(j,k) = sc
      end do
! new matrix a_{k+1} from a_{k}, and eigenvectors
      do k=1,n
        cs =  c*a(k,i)+s*a(k,j)
        sc = -s*a(k,i)+c*a(k,j)
        a(k,i) = cs
        a(k,j) = sc
        cs =  c*x(k,i)+s*x(k,j)
        sc = -s*x(k,i)+c*x(k,j)
        x(k,i) = cs
        x(k,j) = sc
      end do
      end do
      end do
      end do
      return
      end

and the error in my compiler is

200 format (' Eigenvalues and eigenvectors (Jacobi method) ',/, &
1
Error: Unexpected element '&' in format string at (1)
C:\Users\TOSHIBA\AppData\Local\Temp\~Source7.f:34.13:

' Matrix A')
1
Error: Invalid character in name at (1)
C:\Users\TOSHIBA\AppData\Local\Temp\~Source7.f:18.20:

write (*,200)
1
Error: FORMAT label 200 at (1) not defined
 
Last edited by a moderator:
Physics news on Phys.org
  • #2
Please use CODE tags to enclose program source code. This preserves any spacing found in the original program, which is sometimes important for languages like Fortran.

It's also not clear which version of Fortran your compiler is using. Is it Fortran 77, Fortran 90/95, something else?

In any event, the FORMAT statement labeled 200 needs attention. Try eliminating the '&' at the end and put it instead in column 6 of the next line.
 
  • Like
Likes alui
  • #3
Code tags inserted... :-)
 
  • #4
Instead of this:
Code:
write (*,200)
  .
  .
  .
200     format (' Eigenvalues and eigenvectors (Jacobi method) ',/, &
            ' Matrix A')
I would dispense with the format statement altogether, and write something like this:
Code:
 write (*, *) ' Eigenvalues and eigenvectors (Jacobi method)  Matrix A'

The compiler is probably complaining about the line being broken up with an incorrect continuation character being used. If all you're doing is printing a string of characters, I don't see any advantage in using a format statement.

Here's a link to a page that discusses continuation lines in Fortran: http://fortranwiki.org/fortran/show/Continuation+lines
 
  • Like
Likes alui
  • #5
SteamKing said:
Please use CODE tags to enclose program source code. This preserves any spacing found in the original program, which is sometimes important for languages like Fortran.

It's also not clear which version of Fortran your compiler is using. Is it Fortran 77, Fortran 90/95, something else?

In any event, the FORMAT statement labeled 200 needs attention. Try eliminating the '&' at the end and put it instead in column 6 of the next line.

SteamKing said:
Please use CODE tags to enclose program source code. This preserves any spacing found in the original program, which is sometimes important for languages like Fortran.

It's also not clear which version of Fortran your compiler is using. Is it Fortran 77, Fortran 90/95, something else?

In any event, the FORMAT statement labeled 200 needs attention. Try eliminating the '&' at the end and put it instead in column 6 of the next line.

Thank you very much :)
 
  • #6
Mark44 said:
Instead of this:
Code:
write (*,200)
  .
  .
  .
200     format (' Eigenvalues and eigenvectors (Jacobi method) ',/, &
            ' Matrix A')
I would dispense with the format statement altogether, and write something like this:
Code:
 write (*, *) ' Eigenvalues and eigenvectors (Jacobi method)  Matrix A'

The compiler is probably complaining about the line being broken up with an incorrect continuation character being used. If all you're doing is printing a string of characters, I don't see any advantage in using a format statement.

Here's a link to a page that discusses continuation lines in Fortran: http://fortranwiki.org/fortran/show/Continuation lines

Thank you very much :)
 

FAQ: Eigenvalues and eigenvectors of a real symmetric matrix in Fortran

What are eigenvalues and eigenvectors of a real symmetric matrix?

Eigenvalues and eigenvectors are mathematical concepts that are used to study linear transformations. In the context of a real symmetric matrix, eigenvalues are the scalars that represent the scale factor of the eigenvectors. In simpler terms, eigenvalues and eigenvectors help us understand how a matrix transforms a vector.

How are eigenvalues and eigenvectors calculated in Fortran?

In Fortran, the eigenvalues and eigenvectors of a real symmetric matrix can be calculated using the subroutine DSYEV from the LAPACK library. This subroutine uses the QR algorithm to compute the eigenvalues and eigenvectors of a matrix.

What is the significance of a real symmetric matrix in Fortran?

A real symmetric matrix is important in Fortran because it has a special property where its eigenvalues are always real numbers. This makes it easier to calculate and analyze the eigenvalues and eigenvectors using Fortran subroutines. Additionally, real symmetric matrices have many applications in physics, engineering, and computer science.

Can a real symmetric matrix have complex eigenvalues?

No, a real symmetric matrix can only have real eigenvalues. This is because the complex eigenvalues would have to come in complex conjugate pairs, which would violate the symmetry of the matrix. Therefore, a real symmetric matrix always has real eigenvalues and orthogonal eigenvectors.

How are eigenvalues and eigenvectors used in real-world applications?

The concept of eigenvalues and eigenvectors has various applications in real-world problems, such as image processing, data compression, and machine learning. They are also used in physics to study the behavior of systems, in engineering to analyze structures and vibrations, and in computer science to solve linear systems of equations.

Similar threads

Replies
7
Views
2K
Replies
2
Views
6K
Replies
1
Views
2K
Replies
18
Views
3K
Replies
7
Views
1K
Replies
5
Views
779
Replies
10
Views
2K
Replies
1
Views
2K
Replies
5
Views
2K
Back
Top