On a Control Lyapunov Function based Anytime Algorithm for Control of Nonlinear Processes

We present an algorithm to calculate control inputs when available processing resources are time-varying. The basic idea is to calculate the control input to decrease a Lyapunov function value as compared to as many time steps in the past as allowed by the system to be controlled and the available processing resources. We analyze the stability of the resulting closed loop system using stochastic Lyapunov functions and indicate, through numerical simulations, that the performance gains obtained can be significant.