Generalized Value Iteration for Nonaffine Adaptive Tracking Control with Application to a Wastewater Treatment System

In this paper, the generalized value iteration algorithm is utilized to design the tracking controller for the nonaffine wastewater treatment plant. First, the algorithm is initialized by a positive define function to guarantee that the cost function sequence is monotonically nonincreasing. Then, the updating rules of the cost function and the tracking control policy are given under the generalized value iteration framework. Moreover, the model network is constructed to solve the steady control and approximate the system state. The critic and action networks are introduced to approximate the iterative cost function and the iterative tracking control, respectively. Finally, the experimental simulation is provided to validate the feasibility of the present strategy. It is highlighted that the evolving policies during the iterative process are adopted to regulate the concentrations of dissolved oxygen and nitrate nitrogen.