Lyapunov Design of Repetitive Learning Control in Network Management Processes

This paper presents a mathematical framework for learning control based on a feedforward term developed to solve a general control problem in the presence of unknown linear dynamics with a known period. Since the learning based feedforward term is generated from a straightforward Lyapunov like stability analysis, the control design can be extended to other Lyapunov based design techniques to derive hybrid control schemes using learning processes based to compensate for the periodic dynamics and other Lyapunov approaches (adaptive feedforward terms) to compensate for non-periodic dynamics. A hybrid adaptive/learning control scheme is derived to achieve global asymptotic queue length tracking in a TCP/IP network.