Control System based on Reinforcement Learning applied to a Klatt-Engell Reactor

A control system based in Reinforcement Learning (RL) is developed to control a two inputs-two outputs (TITO) system, which is a Klatt-Engell reactor. In the design of the control system, it will be based on a learning method called Model Free Reinforcement Learning (MLFC). Firstly, it will show the model math of the plant to build the nonlinear state equation. Then, it will design the controller based on two sub-controllers MLFC, selecting the characteristics of the states, actions, rewards and other parameters of RL. Finally, it will perform the simulations of the control system for references functions, stepped variables and sine.