Scribd
Upload
187 views2 pages

Python Code For Gauss Method

This Fortran program solves a system of linear equations using Gaussian elimination. It declares matrices to store the coefficients and solutions. It loads the coefficients into the matrix according to the given difference equation. It performs Gaussian elimination to find the unique solution, normalizing all pivots to 1. It writes the solutions to a file and prints them. Subroutines are used to swap rows, strike out elements below the pivot, and normalize pivots during the elimination process.

Uploaded by

Himanshu Gaur
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
Download as pdf or txt
187 views2 pages

Python Code For Gauss Method

This Fortran program solves a system of linear equations using Gaussian elimination. It declares matrices to store the coefficients and solutions. It loads the coefficients into the matrix according to the given difference equation. It performs Gaussian elimination to find the unique solution, normalizing all pivots to 1. It writes the solutions to a file and prints them. Subroutines are used to swap rows, strike out elements below the pivot, and normalize pivots during the elimination process.

Uploaded by

Himanshu Gaur
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 2

File: /home/himanshu/fortran/gauss.

f95 Page 1 of 2

! program fo ques 4, in this i will obtain a diagonal matrix instead of a upper


triangular matrix, i will also normalise all the pivots to 1, so that we can read
the last column as solution

program gauss
implicit none
real(8), dimension(41,42) :: a,b ! declaring matrix a to store all elements and b
for swapping
integer :: i,j,k
logical :: lo;lo=.true.

a(1:41,1:42)=0
a(1,1)=1
a(1,42)=-0.5
a(41,41)=1
a(41,42)=log(2.0)
do i=2,40
a(i,i+1)=(0.025**(-2))+4*((1+0.025*(i-1))*0.025)**(-1)
a(i,i)=-2*(0.025)**(-2)-4*((1+0.025*(i-1))*0.025)**(-1)-2*(1+0.025*(i-1))**(-2)
a(i,i-1)=(0.025)**(-2)
end do ! assignment to all matrix elements according to the given difference
equation

iloop: do i=1,41
! finding pivot
jloop: do j=i,41
if(a(j,j).ne.0) then
exit jloop
end if
end do jloop ! if non zero value is found in the ith column then loop
exits, and value of j is not incremented to 42

if (j.eq.42) then
print*,'no solution'
exit iloop
end if ! in above loop if j reaches 42 it means no non zero element was
found, hence unique sol. is not possible

if(j.ne.i) then
call replace(a,b,i,j) ! calling a subroutine to swap the ith row
with the row having nonzero element
end if

call strike(a,i) ! makes all the elements in a ith column zero other than
pivot

do k=1,41 ! this and the next are extra steps, they normalise all
the pivots to 1,
if(a(k,k).eq.0) then
lo=.false.
print*,"error"
end if
end do ! checking if any pivots are zero, so we don't divide by 0 while
normalising.

if(lo) then
call norm(a)
end if
end do iloop

do i=1,41
write(11,*) (1+0.025*(i-1)), a(i,42) ! all the answers are written to fort.
11
print*, (1+0.025*(i-1)), a(i,42)
end do
File: /home/himanshu/fortran/gauss.f95 Page 2 of 2

end program

subroutine replace(a,b,i,j) ! subroutine to swap


real(8), dimension(41,42) :: a,b
integer i,j
b(1,1:42)=a(i,1:42)
a(i,1:42)=a(j,1:42)
a(j,1:42)=b(1,1:42)
end subroutine

subroutine strike(a,i) ! subroutine to make all elements zero other than pivot in
the given column
real(8), dimension(41,42) :: a
integer i,j
do j=1,41
if (i.ne.j) then
a(j,1:42)=a(j,1:42)-(a(j,i)/a(i,i))*a(i,1:42)
end if
end do
end subroutine

subroutine norm(a) ! subroutine to normalise al pivots


real(8), dimension(41,42) :: a
integer i
do i=1,41
a(i,1:42)=a(i,1:42)/a(i,i)
end do
end subroutine

You might also like