An Alternative Dual Control Approach for Auto-Optimization in Uncertain Environments

This paper investigates the problem of dual control for exploitation and exploration (DCEE) for auto-optimization in uncertain environments. Different from existing adaptive control methods, an exploration effect is additionally considered in the DCEE framework. By providing an alternative dual control approach, the system output can be driven to the estimated nominal value and the uncertainty of the predicted optimal operational condition can be reduced. The convergence analysis of optimality tracking process of general linear systems is established using the DCEE framework. Simulation results are provided to verify the effectiveness of the developed DCEE.