Research
|
 |
A Reinforcement Learning Model for Industrial Filling Process
Control
Ömer Sabri Emeksiz,
Mert Eren Ağcabay,
Engin
Maşazade
Suat Selim,
Taner Boysan
IEEE ICECS, 2023
This paper proposes a reinforcement learning-based industrial filling
process to address limitations of PID controllers in practical cases, such
as discrete action spaces, residual material overflow, and material
properties affecting flow. By recording episodes, defining state-action
pairs, and using rewards and penalties, the Monte Carlo method optimizes
actions to minimize filling duration.
|
|