Mihai,

> just to clear things up in my mind: the way to go from the KKT system (eqns
> 2.8-2.10 in your paper) for the cost function J to a single discrete
> bilinear form (for implementation) is to re-cast the original system into
> one that looks like equation (4.1) and then apply Gauss-Newton to this
> formulation (eqn 5.1-5.2)?

That's one way. What I meant to say is that you should start by writing your 
whole system of equations in a single semilinear form. For example, (2.7) has 
that form -- and if you choose your test functions appropriately, then (2.7) 
is equivalent to either (2.8), (2.9) or (2.10). From there you can decide 
whether you first want to optimize as I do, or first want to discretize. You 
can also decide whether you want a Newton or Gauss-Newton method. But you 
should start from a single semilinear form for the entire set of equations.

By the way, you can find a different viewpoint on this here:
  http://www.dealii.org/7.0.0/doxygen/deal.II/group__vector__valued.html
in the "Philosophy" section.

Best
 W.


-------------------------------------------------------------------------
Wolfgang Bangerth                email:            [email protected]
                                 www: http://www.math.tamu.edu/~bangerth/
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to