Approximately optimal nonlinear stabilization with preservation of the Lyapunov function property

We present an approximate optimization approach to the computation of stabilizing feedback laws using a partitioning of the state space and a corresponding approximation of the optimal value function of the problem. By including the discretization errors into the optimal control formulation we are able to compute approximate optimal value functions which preserve the Lyapunov function property and corresponding optimally stabilizing feedback laws which are constant on each partition element. The actual computation uses efficient graph theoretic algorithms.