Feature request : Passing additional parameters to cost and gradient function.

[pirpyn]

Lets have the following example :

The cost function is f and his gradient is g. x is the optimization variable, real array of size 2. Let y,z be real array of same dimensions ( > 1 )

f = 0.5_wp*sum( ( x(1) + y*x(2) - z )**2 )

g(1) = sum(    x(1) + y*x(2) - z  )
g(2) = sum( y*(x(1) + y*x(2) - z ))

How can we pass y and z to f and g ? Is it already implemented ?

[jacobwilliams]

Yes, you can do that in a couple of different ways. In the example slsqp_test.f90, you can see how the functions are contained within the calling routine. In that case, you could just declare your y and z in the calling routine and then the functions would have access to it.

Alternately, you can extend the slsqp_solver class and add whatever data you want. Something like:

  type,extends(slsqp_solver) :: mysolver
    real(wp) :: y,z
  end type mysolver

Then in your function:

  subroutine func(me,x,f,c)
  implicit none
  class(slsqp_solver),intent(inout) :: me
  real(wp),dimension(:),intent(in)  :: x  
  real(wp),intent(out) :: f  
  real(wp),dimension(:),intent(out) :: c 

  select type (me)
  class is (mysolver)
    f = 0.5_wp*sum( ( x(1) + me%y*x(2) - me%z )**2 )

    ! ... etc

  end select 

  end subroutine func

[pirpyn]

extends is a nice way to do it. Thank you.