Prodromos Daoutidis
University of Minnesota, MN·Engineering
Chemical engineers make decisions with mathematical optimization: finding the best design or operating plan that meets many constraints. Problems like running chemical plants on renewable power, designing resilient supply chains, or operating facilities sustainably involve many processes on different time and length scales.
- Fields
Engineering
Computer Science
Mathematics & Statistics
- Topics
mathematical optimization
nonlinear programming
algorithm selection
- Grants
- 1 known grant, through Aug 2027
What they work on
- Develop a graph neural network framework that represents general nonlinear optimization problems as graphs capturing detailed information about their variables and constraints.
- Use geometric deep learning, machine learning for graph-shaped data, to automatically select the best solution strategy and tune solvers for nonlinear and mixed integer nonlinear problems in process systems engineering.
- Apply explainable AI methods to reveal how features of an optimization problem relate to solver performance, to detect bottlenecks and guide problem reformulation or algorithm improvements.
More about the research
That makes the optimization problems large and complex. Many algorithms exist to solve them, but they are hard and slow to set up, and it is rarely clear in advance which one suits a given problem.
Why it matters
Automatic algorithm selection could let industry engineers and academic researchers solve hard design and operation problems faster. Examples include chemical plants tied to renewable energy and resilient supply chains. Released as open-source software, the framework could also show where current models or solvers fall short and guide improvements.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Haipeng Chen
William & Mary, VA·Computer Science
Develop physics-inspired, task-aware graph neural networks that build problem structure into reinforcement learning, so it can handle badly shaped or implicit objectives in combinatorial optimization.
Philip Yu
University of Illinois Chicago, IL·Computer Science
Integrate open-source anomaly detection tools for tabular, graph and time-series data into a single OpenAD ecosystem with a standardized development environment for building and evaluating models.
Shuai Zhang
New Jersey Institute of Technology, NJ·Engineering
Develop topology sampling and graph neural network pruning methods for single-layer graph models, trimming network structure and model size so spatial-temporal AIoT data can be analyzed more efficiently.
The school
University of Minnesota: 88 labs on PI Pages, $1.4B a year on research, 722 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.