Please do join us for an exciting part of the Sargent Centre Seminar Series on Tuesday 5th July.
The seminar, followed by a Q&A, will be taking place from 3-4pm in the Sargent Centre RODH 617, with the option to join via Microsoft Teams using the livestream link in the left hand column. There will be refreshments following the seminar in RODH 615, with the opportunity to network and to meet the speaker.
A summary of ‘‘Accelerating global optimization for process design: Reduced-space formulations, relaxations & parallelization’:
Many optimization problems in process engineering are inevitably nonconvex. In process design, this applies for example to flowsheet or superstructure optimization. For nonconvex problems, deterministic global optimization methods are required to guarantee optimal solutions. However, these methods are computationally expensive, and despite considerable progress in both algorithms and software over the past decades, many problems of interest are still intractable. In this seminar, we will discuss three approaches for accelerating global optimization in process design, which have been implemented and tested in the open-source optimization solver MAiNGO. First, factorable reduced-space formulations are introduced as a means for exploiting special problem structure arising, e.g., in flowsheeting problems. These formulations reduce the size of the optimization problem exposed to the optimizer and can thus lead to significant speedups. Second, we present custom relaxations of selected thermodynamic models that occur in many design problems. Third, we briefly touch upon the promise and challenges of parallel computing for deterministic global optimization. Finally, an outlook is given for translating these methods to other problem classes.
About the speaker – Dr Dominik Bongartz:
Dr.-Ing. Dominik Bongartz is a group leader at the Chair for Process Systems Engineering (AVT.SVT) at RWTH Aachen University, and a Theodore von Kármán fellow of RWTH. Currently, he is a visiting researcher at CPSE at Imperial College, working with Prof. Benoit Chachuat. Dominik obtained his PhD from RWTH in 2020 for work on deterministic global flowsheet optimization and power-fo-fuel processes, advised by Prof. Alexander Mitsos. Previously, he obtained a Master’s degree in Mechanical Engineering from MIT in 2014 and a Bachelor’s degree in Mechanical Engineering from RWTH in 2012. His research interests include optimization-based process design, global optimization methods, and processes for sustainable energy conversion. Recently, his work has been recognized with the David W. Smith Jr. Graduate Publication award of the CAST division of AIChE and the Young Researcher Award of the International Society of Global Optimization.
Please do get in contact if you have any queries, and please do share this with any colleagues or peers who may find this of interest.