Question: NLPSolve in Matrix Form - variable problem dimensions

Hello,

I ran into problems while trying to write the input to NLPSolve in the Matrix Form because the problem dimensions depend on the input data. Take the following example:

max f(x)           <- f a nonlinear function of x

st: x₁+x₂+x₃+x₄                     ≤ b₁

                                 x₅+x₆+x₇+x₈ ≤ b₂

       x₁      +x₃     +x₅     +x₇        ≤ b₃

            x₂     +x₄        +x₆      +x₈ ≤ b₄

      x₁                     +x₅               ≤ b₅

         x₂                      +x₆          ≤ b₆

                   x₃                     +x₇       ≤ b₇

                         x₄                +x₈ ≤ b₈

                            x ≥ 0.

Depending on the input data, the dimension of x changes as well as the number of constraints and the structure of each constraint (more or less variables in each). Writing out the matrix of coefficients each time there is a change in the input data is definitely not smart. I chose a somewhat lenghthy example to also support the fact that I want (very much) to avoid manual input.

Is there a way to overcome this issue while using the matrix form or is there another more generalized call to the solver? By a more generalized call I mean a command that will allow me to address the set of constraints via a procedure call, something of the form constr:=proc(a,b) for i to 8 do a[i]  ≤ b[i] end do; end proc; and then just use constr in the solver command.

Many thanks in advance,

Melania