How the RL agent knows that it has to provide 'action' as 'FLOW' because none of my input or observations is flow? What decides the output of action?

https://www.mathworks.com/help/reinforcement-learning/ug/create-simulink-environments-for-reinforcement-learning.html

Answer 1

[Vimal Rathod]

Hi,

The details are given properly in the link which you have sent. In the example, input is a scalar variable which indicates the amount of flow (Which is Action) and observation is a 1x3 vector which describes the rise or fall in water level.

You could refer to the following link to know what are the values there in observation and action.

https://www.mathworks.com/help/reinforcement-learning/ug/create-simulink-environments-for-reinforcement-learning.html#WaterTankEnvironmentModelExample-1

To know more about rlNumericSpec which describes a datatype having a range of numeric values refer the following link:

https://www.mathworks.com/help/reinforcement-learning/ref/rl.util.rlnumericspec.html

Hope this helps!

[kaustubh kulkarni]

But vimal action is output,not input...how the RL block understand to calculate FLOW as output with height as input

[kaustubh kulkarni]

With height as only input how block interprets to calculate only volume flow, why not area or other