NAG Library Manual, Mark 30.2
Interfaces:  FL   CL   CPP   AD 

NAG FL Interface Introduction
Example description
!   E04FCF Example Program Text
!   Mark 30.2 Release. NAG Copyright 2024.
    Module e04fcfe_mod

!     E04FCF Example Program Module:
!            Parameters and User-defined Routines

!     .. Use Statements ..
      Use nag_library, Only: nag_wp
!     .. Implicit None Statement ..
      Implicit None
!     .. Accessibility Statements ..
      Private
      Public                           :: lsqfun, lsqgrd, lsqmon
!     .. Parameters ..
      Real (Kind=nag_wp), Parameter    :: one = 1.0_nag_wp
      Real (Kind=nag_wp), Parameter    :: two = 2.0_nag_wp
      Real (Kind=nag_wp), Parameter    :: zero = 0.0_nag_wp
      Integer, Parameter               :: inc1 = 1
      Integer, Parameter, Public       :: liw = 1, m = 15, n = 3, nin = 5,     &
                                          nout = 6, nt = 3
      Integer, Parameter, Public       :: ldfjac = m
      Integer, Parameter, Public       :: ldv = n
      Integer, Parameter, Public       :: lw = 6*n + m*n + 2*m + n*(n-1)/2
      Character (1), Parameter         :: trans = 'T'
!     .. Local Arrays ..
      Real (Kind=nag_wp), Public, Save :: t(m,nt), y(m)
    Contains
      Subroutine lsqgrd(m,n,fvec,fjac,ldfjac,g)
!       Routine to evaluate gradient of the sum of squares

!       .. Use Statements ..
        Use nag_library, Only: dgemv
!       .. Scalar Arguments ..
        Integer, Intent (In)           :: ldfjac, m, n
!       .. Array Arguments ..
        Real (Kind=nag_wp), Intent (In) :: fjac(ldfjac,n), fvec(m)
        Real (Kind=nag_wp), Intent (Out) :: g(n)
!       .. Executable Statements ..
!       The NAG name equivalent of dgemv is f06paf
        Call dgemv(trans,m,n,one,fjac,ldfjac,fvec,inc1,zero,g,inc1)

        g(1:n) = two*g(1:n)

        Return

      End Subroutine lsqgrd
      Subroutine lsqfun(iflag,m,n,xc,fvec,iw,liw,w,lw)

!       Routine to evaluate the residuals

!       .. Scalar Arguments ..
        Integer, Intent (Inout)        :: iflag
        Integer, Intent (In)           :: liw, lw, m, n
!       .. Array Arguments ..
        Real (Kind=nag_wp), Intent (Out) :: fvec(m)
        Real (Kind=nag_wp), Intent (Inout) :: w(lw)
        Real (Kind=nag_wp), Intent (In) :: xc(n)
        Integer, Intent (Inout)        :: iw(liw)
!       .. Executable Statements ..
        fvec(1:m) = xc(1) + t(1:m,1)/(xc(2)*t(1:m,2)+xc(3)*t(1:m,3)) - y(1:m)

        Return

      End Subroutine lsqfun
      Subroutine lsqmon(m,n,xc,fvec,fjac,ldfjac,s,igrade,niter,nf,iw,liw,w,lw)
!       Monitoring routine

!       .. Use Statements ..
        Use nag_library, Only: ddot
!       .. Parameters ..
        Integer, Parameter             :: ndec = 3
!       .. Scalar Arguments ..
        Integer, Intent (In)           :: igrade, ldfjac, liw, lw, m, n, nf,   &
                                          niter
!       .. Array Arguments ..
        Real (Kind=nag_wp), Intent (In) :: fjac(ldfjac,n), fvec(m), s(n),      &
                                          xc(n)
        Real (Kind=nag_wp), Intent (Inout) :: w(lw)
        Integer, Intent (Inout)        :: iw(liw)
!       .. Local Scalars ..
        Real (Kind=nag_wp)             :: fsumsq, gtg
        Integer                        :: j
!       .. Local Arrays ..
        Real (Kind=nag_wp)             :: g(ndec)
!       .. Executable Statements ..
!       The NAG name equivalent of ddot is f06eaf
        fsumsq = ddot(m,fvec,inc1,fvec,inc1)

        Call lsqgrd(m,n,fvec,fjac,ldfjac,g)

        gtg = ddot(n,g,inc1,g,inc1)

        Write (nout,*)
        Write (nout,*)                                                         &
          '  Itn      F evals        SUMSQ             GTG        Grade'
        Write (nout,99999) niter, nf, fsumsq, gtg, igrade
        Write (nout,*)
        Write (nout,*)                                                         &
          '       X                    G           Singular values'

        Write (nout,99998)(xc(j),g(j),s(j),j=1,n)

        Return

99999   Format (1X,I4,6X,I5,6X,1P,E13.5,6X,1P,E9.1,6X,I3)
99998   Format (1X,1P,E13.5,10X,1P,E9.1,10X,1P,E9.1)
      End Subroutine lsqmon
    End Module e04fcfe_mod
    Program e04fcfe

!     E04FCF Example Main Program

!     .. Use Statements ..
      Use e04fcfe_mod, Only: ldfjac, ldv, liw, lsqfun, lsqgrd, lsqmon, lw, m,  &
                             n, nin, nout, nt, t, y
      Use nag_library, Only: e04fcf, nag_wp, x02ajf
!     .. Implicit None Statement ..
      Implicit None
!     .. Local Scalars ..
      Real (Kind=nag_wp)               :: eta, fsumsq, stepmx, xtol
      Integer                          :: i, ifail, iprint, maxcal, nf, niter
!     .. Local Arrays ..
      Real (Kind=nag_wp)               :: fjac(m,n), fvec(m), g(n), s(n),      &
                                          v(ldv,n), w(lw), x(n)
      Integer                          :: iw(liw)
!     .. Intrinsic Procedures ..
      Intrinsic                        :: sqrt
!     .. Executable Statements ..
      Write (nout,*) 'E04FCF Example Program Results'

!     Skip heading in data file
      Read (nin,*)

!     Observations of TJ (J = 1, 2, ..., nt) are held in T(I, J)
!     (I = 1, 2, ..., m)

      Do i = 1, m
        Read (nin,*) y(i), t(i,1:nt)
      End Do

!     Set IPRINT to 1 to obtain output from LSQMON at each iteration

      iprint = -1

      maxcal = 400*n
      eta = 0.5_nag_wp
      xtol = 10.0_nag_wp*sqrt(x02ajf())

!     We estimate that the minimum will be within 10 units of the
!     starting point

      stepmx = 10.0_nag_wp

!     Set up the starting point

      x(1:nt) = (/0.5_nag_wp,1.0_nag_wp,1.5_nag_wp/)

      ifail = -1
      Call e04fcf(m,n,lsqfun,lsqmon,iprint,maxcal,eta,xtol,stepmx,x,fsumsq,    &
        fvec,fjac,ldfjac,s,v,ldv,niter,nf,iw,liw,w,lw,ifail)

      Select Case (ifail)
      Case (0,2:)
        Write (nout,*)
        Write (nout,99999) 'On exit, the sum of squares is', fsumsq
        Write (nout,99999) 'at the point', x(1:n)

        Call lsqgrd(m,n,fvec,fjac,ldfjac,g)

        Write (nout,99998) 'The estimated gradient is', g(1:n)
        Write (nout,*) '                           (machine dependent)'
        Write (nout,*) 'and the residuals are'
        Write (nout,99997) fvec(1:m)
      End Select

99999 Format (1X,A,3F12.4)
99998 Format (1X,A,1P,3E12.3)
99997 Format (1X,1P,E9.1)
    End Program e04fcfe