Local minimization algorithms for dynamic programming equations

Dante Kalise (1) , Axel Kröner (2, 3) , Karl Kunisch (4, 1)

1 RICAM - Johann Radon Institute for Computational and Applied Mathematics
2 CMAP - Centre de Mathématiques Appliquées - Ecole Polytechnique
3 Commands - Control, Optimization, Models, Methods and Applications for Nonlinear Dynamical Systems
4 IM - Institut für Mathematik [Graz]

The numerical realization of the dynamic programming principle for continuous-time optimal control leads to nonlinear Hamilton-Jacobi-Bellman equations which require the minimization of a nonlinear mapping over the set of admissible controls. This minimization is often performed by comparison over a finite number of elements of the control set. In this paper we demonstrate the importance of an accurate realization of these minimization problems and propose algorithms by which this can be achieved effectively. The considered class of equations includes nonsmooth control problems with l1-penalization which lead to sparse controls.

Domaines: Optimisation et contrôle [math.OC]; Analyse numérique [math.NA]

Dante Kalise, Axel Kröner, Karl Kunisch. Local minimization algorithms for dynamic programming equations . SIAM Journal on Scientific Computing, 2016, 38 (3). ⟨hal-01120450⟩