6.2 Mathematical and numerical methods employed during the CEMRACS 97:

1. Unconstrained optimization - The function to minimize is continuous differentiable and its gradient can be computed analytically;
2. Augmented Lagrangian;
3. Nonsmooth optimization (based on the works of E. Polak, Berkeley) - Existence of an optimal solution, first order optimal conditions, nonsmooth steepest descent method. Using the particularity of the constraints, we have reduced at half the number of the variables;
4. Multigrid method;
5. Continuous distribution of the sources;