Occupation measures and semidefinite relaxations for optimal control

This thesis details a global method for optimal control of open-loop systems. This is done by relaxing the control problem as a generalized moment problem, which can be solved numerically by a hierarchy of semi-definite relaxations. The approach is first applied to the impulsive control of linear time varying systems, by modeling the controls by a measure. The resulting semi-definite conditions circumvent time discretiziation and related difficulties. By the use of occupation measures, the method is then extended to a class of impulsive non-linear problems. This results in a monotone sequence of lower bounds to the original control problem. Finally, those results are transposed to switched system, by modeling each mode by a corresponding occupation measure. This allows for large computational gains with respect to the classical approach, where the control space is measured.