Numerical investigation of a stability theorem arising from the 2-dimensional analysis of an iterative optimal control algorithm

Because many optimal control problems require solution using iterative procedures they fall naturally in the realm of 2-D systems where the two dimensions are response time horizon and iteration index, respectively. The paper uses this observation to employ 2-D systems theory, in the form of unit memory repetitive process techniques, to analyse local stability and convergence behaviour of a continuous optimal control algorithm based on dynamic system optimisation and parameter estimation. Existing work is extended to incorporate unmatched terminal constraints. Necessary and sufficient conditions for stability are obtained whose evaluation require the solution of a difficult eigenvalue problem. The paper shows how solutions can be achieved using numerical and graphical facilities of MATLAB.