Tank-Level-Control-with-Reinforcment-Learning

This repository is still being updated for detailed explaiation of the system and a batter navigation through the code. The implementation for non-linear CSTR (MIMO) is in process and will be uploaded soon.

This repository demonstrates the Reinforcement Learning TD Q-Learning algorithm to control the level of the tank.

System Information

State - Action

In order to track multiple setpoints, the state for this system is the current setpoint tracking error ( difference between setpoint and height ). Simillarly, the action is the necessary change in the current input.

An Example