Constructs a Tikhonov problem:
Minimize \(J(d) = J_S(F(d)-u_obs) + \frac{1}{\eta} J_D(d-d0) \)
that can be solved with a TaoBasicSolver.
- Parameters
-
| forward | Class defining the map F. See class-level documentation for requirements of F. |
| d0 | Best a-priori guess for the design parameter. |
| u_obs | State parameter to match (i.e. approximately solve F(d)=u_obs) |
| eta | Penalty parameter/Lagrange multiplier. Take eta to zero to impose more regularization to an ill posed problem. |
| designFunctional | The functional \(J_D\) |
| stateFunctional | The functional \(J_S\) |
Definition at line 304 of file IPTaoTikhonovProblem.hh.
References pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_d, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_d0, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_d_diff, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_dGlobal, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_grad, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_grad_design, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_grad_state, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_grid, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_tikhonov_atol, pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_tikhonov_rtol, and pism::inverse::IPTaoTikhonovProblem< ForwardProblem >::m_u_diff.
