Solving Optimization Problems Using Parallel and Grid Computing

Optimization problems involving millions of variables arise in three classes of applications: optimal control, inverse problems and shape optimization. In all these cases the simulation requires the solution of 3D partialdifferential equations, which must be peformed using parallel computing environments. We will present case studies illustrating the state-of-the-art of this field, which is often called “PDE-constrained optimization”, and we will discuss open problems and future research directions.