Past, present and future projects
Flowsheet design with RL
A Reinforcement Learning Approach with Masked Agents for Chemical Process Flowsheet Design
This study presents two innovative masked Reinforcement Learning agents designed for optimizing chemical process flowsheets. A key innovation is the incorporation of masking within the hybrid framework, which utilizes expert input to exclude certain actions from the agent's decision-making process. Through case studies, including simulations with ASPEN Plus®, the efficacy of these agents is demonstrated, revealing their ability to learn and successfully identify flowsheet designs that meet specific process requirements, such as achieving defined product quality.
The code can be found in the research group's repository
Flowsheet design and control with RL
An integrated reinforcement learning framework for simultaneous generation, design, and control of chemical process flowsheets
This study presents a Reinforcement Learning (RL) approach for the synthesis, design, and control of chemical process flowsheets (CPFs). The RL framework uses inlet streams and unit operations to build and evaluate CPFs, while leveraging surrogate models, particularly Neural Networks, to speed up learning and reduce dependence on mechanistic dynamic models. Case studies demonstrate the framework's ability to maintain dynamic operability, adhere to constraints, and generate viable, economically attractive CPFs.
The code can be found in the research group's repository