Advanced Policy Iteration with Novel Initialization for Online Nonlinear Optimal Regulation

In this paper, a novel online control framework is constructed to address the optimal regulation problem for nonlinear systems. Produced by policy iteration, the proposed algorithm employs iterative policies to control the system. To initialize the policy iteration process, the value iteration method is used to obtain an admissible control policy. It is shown that the stability can be ensured via different admissible iterative policies. Finally, a simulation example is presented to demonstrate the performance of the developed scheme.