2025 AIChE Annual Meeting

Session: 10B: AI/ML and Data-Driven Modeling for Control I

This session welcomes contributions that highlight the application of AI/ML and data-driven technologies for modeling with control-relevant applications, as well as theory for AI/ML and data-driven methods as applied to modeling in control-relevant applications.

Chair

Davood Pourkargar, Kansas State University

Co-Chairs

Dhan Lord Fortela, University of Louisiana at Lafayette
Akhilesh Jain, Sparkcognition
Manish Aggarwal, AspenTech Corporation

Presentations

08:00 AM

(259a) Continuous Learning of the Value Function Utilizing Deep Reinforcement Learning and Its Use As the Objective in Model Predictive Control
Daniel Beahr, Elijah Hedrick, Debangsu Bhattacharyya

08:18 AM

(259b) Control-Informed Reinforcement Learning
Maximilian Bloor, Akhil Ahmed, Niki Kotecha, Mehmet Mercangöz, Calvin Tsay, Antonio del Rio Chanona

08:36 AM

(259c) Automated Discovery of Diverse Disturbance Scenarios Via Combinatorial Multi-Arm Bandits and Time-Series Diffusion Models: Application to Building Control Systems
Wei-Ting Tang, Abraham Vinod, François Germain, Chris Laughman, Ankush Chakrabarty, Joel Paulson

08:54 AM

(259d) Leveraging Large Language Models for Dynamic Fault Recovery Via Finite State Machine Traversal
Javal Vyas, Milapji Gill, Felix Gehlhoff, Mehmet Mercangöz

09:12 AM

(259e) State-Space Neural Network Architecture for Nonlinear Dynamical System Modeling
Aswin Chandrasekar, Prashant Mhaskar

09:30 AM

(259f) Physics-Informed Model-Based Policy Optimization of Koopman Economic NMPC Policies
Daniel Mayfrank, Mehmet Velioglu, Alexander Mitsos, Manuel Dahmen

09:48 AM

(259g) Koopman System Identification with Noise Modeling
Moritz Woelk, Wentao Tang

10:06 AM

(259h) Iterative Model Identification Scheme Via Gaussian Process-Based NMPC for Batch Processes
Tai Xuan Tan, Alexander Mitsos, Eike Cramer