Dr. John Paul Watson
Senior Research Scientist
Lawrence Livermore National Laboratory (LLNL)
Friday, September 2, 2022
2-3pm Tickle 500
https://tennessee.zoom.us/j/96327677336
Seminar abstract:
There are various modeling paradigms available for optimization problems under uncertainty. In this talk, we focus on stochastic programming – in which uncertainty is represented by a discrete set of scenarios. We describe fundamentals of stochastic programming models, and discuss how the paradigm can be practically used to express real-world optimization models where uncertainty is a central issue. Beyond modeling, solution approaches are a major challenge for practical deployment of stochastic programming models. To support both modeling and scalable solution, we introduce the mpi-sppy Python-based library for stochastic programming. We discuss challenges associated with solving real-world stochastic programs and how mpi-sppy addresses those challenges. Results using high-performance computing platforms are presented.
Bio-sketch:
Dr. Jean-Paul Watson is a Senior Research Scientist at Lawrence Livermore National Laboratory (LLNL) in Livermore, California. In addition to his position in the Center for Applied Scientific Computing (CASC), he is the Associate Program Leader for Climate Resilience in the Cyber and Infrastructure Resilience (CIR) Program within the Global Security Directorate at LLNL. He has worked in the US national lab system (previously at Sandia National Laboratories) for nearly 20 years, primarily leading projects in foundational and applied mathematical optimization. His main research interests involve optimization under uncertainty and applications to critical infrastructure operations, planning, and resilience. He is co-inventor of the widely used Pyomo (www.pyomo.org) Python-based library for mathematical optimization, which received the R&D 100 award in 2016. He has also been awarded the INFORMS Computing Society prize and was an INFORMS Edelman finalist.