Optimal Control Mesh

We present a novel algorithm for optimal control of nonlinear systems, which creates control function over a mesh on a region of interest. The algorithm presented in this paper is an alternative to a set-oriented approach and subdivision algorithm for optimal control. The main contribution of this paper is error estimation for a found solution. We show on two dimensional and three dimensional problems, that this new algorithm is faster than a subdivision algorithm. In comparison with a set-oriented approach, the new algorithm keeps the same advantages as the subdivision algorithm which include a smaller memory foot-print of the final solution, no need for discretization and knowledge when to stop increasing the mesh size.