On-line optimization of fedbatch bioreactors by adaptive extremum seeking control

Cougnon, Pascal; Dochain, Denis; Guay, Martin; Perrier, Michel

doi:10.1016/j.jprocont.2011.05.004

Cited by 55 publications

(20 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This model is applicable to wide range of other systems where cell or microorganism growth is important, including wastewater treatment [22] and the gut microbiome [23]. Such systems can be especially difficult to control because they often are equipped with minimal online sensors [24], limiting the effectiveness of classical control techniques that are hampered by infrequent or delayed system measurements [20,25].…”

Section: Introductionmentioning

confidence: 99%

Deep reinforcement learning for the control of microbial co-cultures in bioreactors

Treloar

Fedorec

Ingalls

et al. 2018

Preprint

View full text Add to dashboard Cite

AbstractMulti-species microbial communities are widespread in natural ecosystems. When employed for biomanufacturing, engineered synthetic communities have shown increased productivity (in comparison with pure cultures) and allow for the reduction of metabolic load by compartmentalising bioprocesses between multiple sub-populations. Despite these benefits, co-cultures are rarely used in practice because control over the constituent species of an assembled community has proven challenging. Here we demonstrate, in silico, the efficacy of an approach from artificial intelligence – reinforcement learning – in the control of co-cultures within continuous bioreactors. We confirm that feedback via reinforcement learning can be used to maintain populations at target levels, and that model-free performance with bang-bang control can outperform traditional proportional integral controller with continuous control, when faced with infrequent sampling. Further, we demonstrate that a satisfactory control policy can be learned in one twenty-four hour experiment, by running five bioreactors in parallel. Finally, we show that reinforcement learning can directly optimise the output of a co-culture bioprocess. Overall, reinforcement learning is a promising technique for the control of microbial communities.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep reinforcement learning for the control of microbial co-cultures in bioreactors

Treloar

Fedorec

Ingalls

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…In this framework, proposes a combination scheme of different types of gradient‐based optimization methods with a parameter estimation algorithm. An intensive research activity has been devoted to adaptive model‐based ESC schemes applied to (bio)‐reaction systems (e.g., . The adaptive extremum schemes are based on parameter learning laws for the estimation of the unknown parameters and a dither signal to ensure the convergence to a neighborhood of the optimal value .…”

Section: Introductionmentioning

confidence: 99%

Robust sliding mode‐based extremum‐seeking controller for reaction systems via uncertainty estimation approach

Lara-Cisneros

Femat

Dochain

2017

Intl J Robust & Nonlinear

Self Cite

View full text Add to dashboard Cite

Summary This paper deals with the design of a robust sliding mode‐based extremum‐seeking controller aimed at the online optimization of a class of uncertain reaction systems. The design methodology is based on an input–output linearizing method with variable‐structure feedback, such that the closed‐loop system converges to a neighborhood of the optimal set point with sliding mode motion. In contrast with previous extremum‐seeking control algorithms, the control scheme includes a dynamic modelling‐error estimator to compensate for unknown terms related with model uncertainties and unmeasured disturbances. The proposed online optimization scheme does not make use of a dither signal or a gradient‐based optimization algorithm. Practical stabilizability for the closed‐loop system around to the unknown optimal set point is analyzed. Numerical experiments for two nonlinear processes illustrate the effectiveness of the proposed robust control scheme. Copyright © 2017 John Wiley & Sons, Ltd.

show abstract

“…Model predictive control (MPC) is the most popular scheme that has been considered due to the simplicity of formulating a nonlinear optimization problem, which can be normally solved by standard numerical methods [6][7][8]. Nevertheless, recently, extremum seeking control (ESC) has emerged as a promising on-line optimization strategy [9][10][11][12]. The main advantage of ESC over MPC is that a formal and rigorous analysis can be carried out in order to assure the stability of the closed-loop dynamics and the convergence of the algorithm to the optimal solution.…”

Section: Introductionmentioning

confidence: 99%

On-line heuristic optimization strategy to maximize the hydrogen production rate in a continuous stirred tank reactor

Ramírez-Morales

Zúñiga

Buitrón

2015

Process Biochemistry

View full text Add to dashboard Cite

On-line optimization of fedbatch bioreactors by adaptive extremum seeking control

Cited by 55 publications

References 11 publications

Deep reinforcement learning for the control of microbial co-cultures in bioreactors

Deep reinforcement learning for the control of microbial co-cultures in bioreactors

Robust sliding mode‐based extremum‐seeking controller for reaction systems via uncertainty estimation approach

On-line heuristic optimization strategy to maximize the hydrogen production rate in a continuous stirred tank reactor

Contact Info

Product

Resources

About