Research Themes
Our group works across several interconnected research themes. Click "Selected Papers" to explore key publications in each area.
Bayesian Optimization & Data-Driven Optimization
Many engineering optimization problems can be described as 'costly' black-box problems, where the number of function evaluations is constrained. Our research develops efficient optimization methods that rely solely on function evaluations — from surrogate-based approaches such as Bayesian optimization to direct derivative-free methods. We also explore human-in-the-loop Bayesian optimization, enabling domain experts such as engineers and chemists to inform the decision-making process. This spans applications including experimental design, catalyst discovery, and bioprocess optimization.
Selected Papers ▼- A Guide to Bayesian Optimization in Bioprocess Engineering
M. Siska, E. Pajak, K. Rosenthal, EA del Rio Chanona, E. von Lieres, LM Helleckes
Biotechnology and Bioengineering, 2026
[Link] - Human-algorithm collaborative Bayesian optimization for engineering systems
T. Savage, EA del Rio Chanona
Computers & Chemical Engineering, 2024
[Link] - Machine learning-assisted discovery of flow reactor designs
T. Savage, N. Basha, J. McDonough, J. Krassowski, O. Matar, EA del Rio Chanona
Nature Chemical Engineering, 2024
[Link] - Real-time optimization meets Bayesian optimization and derivative-free optimization: A tale of modifier adaptation
EA del Rio Chanona, P. Petsagkourakis, E. Bradford, JE Alves Graciano, B. Chachuat
Computers & Chemical Engineering, 2021
[Link] - Data-driven optimization for process systems engineering applications
D. van de Berg, T. Savage, P. Petsagkourakis, D. Zhang, N. Shah, EA del Rio-Chanona
Chemical Engineering Science, 2022
[Link] - Data-driven coordination of expensive black-boxes
D. van de Berg, P. Petsagkourakis, N. Shah, EA del Rio-Chanona
Computer Aided Chemical Engineering, 2022
[Link]
Reinforcement Learning & Process Control
Reinforcement Learning (RL) is a subfield of AI which trains models to make optimal sequential decisions. Chemical processes are inherently stochastic and dynamic — making RL a natural fit. However, traditional RL is data-hungry and does not consider safety constraints, which is a major drawback for process engineering. Our research designs new RL algorithms that can optimize complex chemical and biochemical processes while respecting safety constraints and requiring fewer data. We also develop advanced model predictive control strategies and hierarchical control architectures that integrate learning-based methods with classical control.
Selected Papers ▼- Control-informed reinforcement learning for chemical processes
M. Bloor, A. Ahmed, N. Kotecha
Industrial & Engineering Chemistry Research, 2025
[Link] - PC-Gym: Benchmark environments for process control problems
M. Bloor, J. Torraca, IO Sandoval, A. Ahmed
Computers & Chemical Engineering, 2025
[Link] - A survey and tutorial of reinforcement learning methods in process systems engineering
M. Bloor, M. Mowbray, EA del Rio Chanona
Computers & Chemical Engineering, 2025
[Link] - Data-driven Koopman MPC using mixed stochastic-deterministic tubes
Z. Zhong, EA del Rio-Chanona
Journal of Process Control, 2025
[Link] - Reinforcement learning meets bioprocess control through behavior cloning: Real-world deployment in an industrial photobioreactor
JD Gil, EA Del Rio Chanona, JL Guzmán, M. Berenguel
Engineering Applications of Artificial Intelligence, 2026
[Link] - Leveraging graph neural networks and multi-agent reinforcement learning for inventory control in supply chains
N. Kotecha, EA del Rio Chanona
Computers & Chemical Engineering, 2025
[Link] - Gaussian Process Q-Learning for Finite-Horizon Markov Decision Processes
M. Bloor, T. Savage, C. Tsay, EA del Rio Chanona, M. Mowbray
Reinforcement Learning Journal, 2025
[Link] - Chance Constrained Policy Optimization for Process Control and Optimization
P. Petsagkourakis, IO Sandoval, E. Bradford, F. Galvanin, D. Zhang, EA del Rio-Chanona
Journal of Process Control, 2022
[Link]
Large Language Models & Generative AI
Our research leverages large language models (LLMs) and generative AI to support decision-making in process systems engineering. We develop multi-agent LLM frameworks for automating sustainable operational decisions, apply chemistry-aware language models to trace material flows across industrial value chains, and build AI agents for tasks such as chemical lab management and process safety. Our goal is to harness the reasoning and generative capabilities of LLMs to accelerate engineering workflows, from supply chain analysis to experiment planning.
Selected Papers ▼- Multi-Agent LLMs for Automating Sustainable Operational Decision-Making
E. Pajak, A. Bahamdan, K. Hellgardt, A. del Río-Chanona
Systems and Control Transactions, 2025
[Link] - CarAT: carbon atom tracing across industrial chemical value chains via chemistry language models
E. Pajak, D. Walz, O. Walz, LM Helleckes, K. Hellgardt, A. del Rio Chanona
Green Chemistry, 2026
[Link] - An analysis of multi-agent reinforcement learning for decentralized inventory control systems
M. Mousa, D. van de Berg, N. Kotecha
Computers & Chemical Engineering, 2024
[Link]
Discovery & Molecular Design
We apply machine learning to accelerate scientific discovery, from catalyst design to molecular property prediction. This includes symbolic regression for interpretable model discovery, machine learning interatomic potentials for molecular dynamics, Bayesian optimization for chemical experiments, and deep learning methods for chemical engineering applications. Our work also covers hybrid modelling — combining first-principles models with data-driven approaches to create more accurate representations of complex systems, leveraging Bayesian inference, transfer learning, and multiscale modeling.
Selected Papers ▼- Constraint-Guided Symbolic Regression for Data-Efficient Kinetic Model Discovery
M-C. Servia, IO Sandoval, K. Kuok, K. Hellgardt
arXiv preprint, 2025
[Link] - Simplest mechanism builder algorithm (SIMBA): an automated microkinetic model discovery tool
MÁ de Carvalho Servia, KK Hii, K. Hellgardt, D. Zhang, EA del Rio Chanona
Chemical Science, 2026
[Link] - MolPrice: assessing synthetic accessibility of molecules based on market value
F. Hastedt, K. Hellgardt, S. Yaliraki, D. Zhang, EA del Rio Chanona
Journal of Cheminformatics, 2025
[Link] - Investigating the reliability and interpretability of machine learning frameworks for chemical retrosynthesis
F. Hastedt, RM Bailey, K. Hellgardt, SN Yaliraki, EA del Rio Chanona, D. Zhang
Digital Discovery, 2024
[Link] - Physics-Informed Automated Discovery of Kinetic Models
MA de Carvalho Servia, IO Sandoval
AIChE Annual Meeting, 2024
[Link] - Cross-disciplinary perspectives on the potential for artificial intelligence across chemistry
AM Mroz, AR Basford, F. Hastedt
Chemical Society Reviews, 2025
[Link]
Sustainability & Supply Chain Optimization
We develop optimization frameworks for sustainable chemical value chains and supply chain operations. This involves mathematical modeling, simulation, and optimization techniques to reduce emissions, improve efficiency, and support the transition toward sustainable chemicals manufacturing. Our research spans the reconfiguration of supply chains, design and operations of integrated energy systems, and graph-based machine learning for configuration of chemical value chains — often in collaboration with industry partners such as BASF.
Selected Papers ▼- Distributional reinforcement learning for inventory management in multi-echelon supply chains
G. Wu, MA de Carvalho Servia, M. Mowbray
Digital Chemical Engineering, 2023
[Link] - Hierarchical planning-scheduling-control — Optimality surrogates and derivative-free optimization
D. van de Berg, N. Shah, EA del Rio-Chanona
Computers & Chemical Engineering, 2024
[Link] - CarAT: carbon atom tracing across industrial chemical value chains via chemistry language models
E. Pajak, D. Walz, O. Walz, LM Helleckes, K. Hellgardt, A. del Rio Chanona
Green Chemistry, 2026
[Link] - Leveraging graph neural networks and multi-agent reinforcement learning for inventory control in supply chains
N. Kotecha, EA del Rio Chanona
Computers & Chemical Engineering, 2025
[Link] - Data-driven coordination of subproblems in enterprise-wide optimization under uncertainty
D. van de Berg, P. Petsagkourakis, N. Shah, EA del Rio-Chanona
AIChE Journal, 2023
All Publications
A comprehensive collection of selected publications from across all of our research themes — spanning optimization, control, AI, molecular design, and sustainability.
Selected Papers ▼- CarAT: carbon atom tracing across industrial chemical value chains via chemistry language models
E. Pajak, D. Walz, O. Walz, LM Helleckes, K. Hellgardt, A. del Rio Chanona
Green Chemistry, 2026
[Link] - Simplest mechanism builder algorithm (SIMBA): an automated microkinetic model discovery tool
MÁ de Carvalho Servia, KK Hii, K. Hellgardt, D. Zhang, EA del Rio Chanona
Chemical Science, 2026
[Link] - A Guide to Bayesian Optimization in Bioprocess Engineering
M. Siska, E. Pajak, K. Rosenthal, EA del Rio Chanona, E. von Lieres, LM Helleckes
Biotechnology and Bioengineering, 2026
[Link] - Reinforcement learning meets bioprocess control through behavior cloning: Real-world deployment in an industrial photobioreactor
JD Gil, EA Del Rio Chanona, JL Guzmán, M. Berenguel
Engineering Applications of Artificial Intelligence, 2026
[Link] - Constraint-Guided Symbolic Regression for Data-Efficient Kinetic Model Discovery
M-C. Servia, IO Sandoval, K. Kuok, K. Hellgardt
arXiv preprint, 2025
[Link] - Cross-disciplinary perspectives on the potential for artificial intelligence across chemistry
AM Mroz, AR Basford, F. Hastedt
Chemical Society Reviews, 2025
[Link] - Multi-Agent LLMs for Automating Sustainable Operational Decision-Making
E. Pajak, A. Bahamdan, K. Hellgardt, A. del Río-Chanona
Systems and Control Transactions, 2025
[Link] - Control-informed reinforcement learning for chemical processes
M. Bloor, A. Ahmed, N. Kotecha
Industrial & Engineering Chemistry Research, 2025
[Link] - PC-Gym: Benchmark environments for process control problems
M. Bloor, J. Torraca, IO Sandoval, A. Ahmed
Computers & Chemical Engineering, 2025
[Link] - A survey and tutorial of reinforcement learning methods in process systems engineering
M. Bloor, M. Mowbray, EA del Rio Chanona
Computers & Chemical Engineering, 2025
[Link] - Data-driven Koopman MPC using mixed stochastic-deterministic tubes
Z. Zhong, EA del Rio-Chanona
Journal of Process Control, 2025
[Link] - MolPrice: assessing synthetic accessibility of molecules based on market value
F. Hastedt, K. Hellgardt, S. Yaliraki, D. Zhang, EA del Rio Chanona
Journal of Cheminformatics, 2025
[Link] - Leveraging graph neural networks and multi-agent reinforcement learning for inventory control in supply chains
N. Kotecha, EA del Rio Chanona
Computers & Chemical Engineering, 2025
[Link] - Gaussian Process Q-Learning for Finite-Horizon Markov Decision Processes
M. Bloor, T. Savage, C. Tsay, EA del Rio Chanona, M. Mowbray
Reinforcement Learning Journal, 2025
[Link] - Human-algorithm collaborative Bayesian optimization for engineering systems
T. Savage, EA del Rio Chanona
Computers & Chemical Engineering, 2024
[Link] - Machine learning-assisted discovery of flow reactor designs
T. Savage, N. Basha, J. McDonough, J. Krassowski, O. Matar, EA del Rio Chanona
Nature Chemical Engineering, 2024
[Link] - An analysis of multi-agent reinforcement learning for decentralized inventory control systems
M. Mousa, D. van de Berg, N. Kotecha
Computers & Chemical Engineering, 2024
[Link] - Physics-Informed Automated Discovery of Kinetic Models
MA de Carvalho Servia, IO Sandoval
AIChE Annual Meeting, 2024
[Link] - Hierarchical planning-scheduling-control — Optimality surrogates and derivative-free optimization
D. van de Berg, N. Shah, EA del Rio-Chanona
Computers & Chemical Engineering, 2024
[Link] - Investigating the reliability and interpretability of machine learning frameworks for chemical retrosynthesis
F. Hastedt, RM Bailey, K. Hellgardt, SN Yaliraki, EA del Rio Chanona, D. Zhang
Digital Discovery, 2024
[Link] - Data-driven coordination of subproblems in enterprise-wide optimization under uncertainty
D. van de Berg, P. Petsagkourakis, N. Shah, EA del Rio-Chanona
AIChE Journal, 2023
[Link] - Distributional reinforcement learning for inventory management in multi-echelon supply chains
G. Wu, MA de Carvalho Servia, M. Mowbray
Digital Chemical Engineering, 2023
[Link] - Data-driven optimization for process systems engineering applications
D. van de Berg, T. Savage, P. Petsagkourakis, D. Zhang, N. Shah, EA del Rio-Chanona
Chemical Engineering Science, 2022
[Link] - Data-driven coordination of expensive black-boxes
D. van de Berg, P. Petsagkourakis, N. Shah, EA del Rio-Chanona
Computer Aided Chemical Engineering, 2022
[Link] - Chance Constrained Policy Optimization for Process Control and Optimization
P. Petsagkourakis, IO Sandoval, E. Bradford, F. Galvanin, D. Zhang, EA del Rio-Chanona
Journal of Process Control, 2022
[Link] - Real-time optimization meets Bayesian optimization and derivative-free optimization: A tale of modifier adaptation
EA del Rio Chanona, P. Petsagkourakis, E. Bradford, JE Alves Graciano, B. Chachuat
Computers & Chemical Engineering, 2021
[Link]