A reinforcement learning controller, pre-trained by imitating a PID controller and fine-tuned online, was deployed for pH regulation in an industrial photobioreactor, with simulation showing about 8% lower integrated absolute error and 54% lower control effort than PID.
, author Guzmán, J.L
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.SY 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Reinforcement learning meets bioprocess control through behaviour cloning: Real-world deployment in an industrial photobioreactor
A reinforcement learning controller, pre-trained by imitating a PID controller and fine-tuned online, was deployed for pH regulation in an industrial photobioreactor, with simulation showing about 8% lower integrated absolute error and 54% lower control effort than PID.