Ostrowski, Herrman's NSF Award Funds Interdisciplinary Optimization

Doctors need to determine a patient's short-term treatment plan before test results come back from the lab. Power grid operators need to decide how to dispense generator assets today without knowing what tomorrow's output or demand will be. Port administrators have to plan freight truck schedules without knowing the exact time each ship will dock.

These are examples of multi-stage stochastic (random) problems: sequences of decisions made over time, where each new decision is based on how a previously uncertain event unfolded. Thousands of decisions like these are made across the energy, logistics, healthcare, finance, and other industries every day.

"Getting these decisions right has enormous practical consequences," said University of Tennessee, Knoxville Department of Industrial and Systems Engineering (ISE) Professor and Associate Department Head James Ostrowski. "Improving our ability to solve them efficiently is fundamental to helping organizations make better decisions in a complex and uncertain world."

Determining the right decision to make after every possible outcome for an uncertain event, even events with binary yes/no outcomes, creates an exponentially increasing number of scenarios over time. Classic computational methods, which individually craft and evaluate each possible scenario, cannot accurately handle such enormous problems.

Quantum computers, on the other hand, are a natural fit.

"Rather than enumerating scenarios one by one," Ostrowski explained, "a quantum circuit can encode all scenarios into a single quantum state simultaneously through a property called superposition."

This spring, Ostrowski and ISE Assistant Professor Rebekah Herrman are jointly embarking on a two-year, $300,000 National Science Foundation (NSF) grant to create quantum computing-based tools that will help researchers and industry engineers quickly determine whether quantum computing can help solve a given two-step uncertainty optimization problem—a vital foundation toward higher-stage scenarios.

"This grant shows the leading role of the ISE department in quantum computing research both at the University of Tennessee and in the nation," said ISE Department Head Mingzhou Jin.

Complementary Computational Strengths

Ostrowski, Herrman, and two Ph.D. students funded by the grant will utilize the world-class quantum computing facilities at Oak Ridge National Laboratory's Quantum Computing User Program. They will also draw on UT's strong interdisciplinary tradition in operations research, computational science, and energy systems engineering, Ostrowski said.

The team's research will harness a hybrid approach, utilizing quantum computation to encode scenario structures and explore the large solution landscapes and classical computation for parameter optimization, solution evaluation, and post-processing.

"Classical and quantum computation have fundamentally different, complementary strengths," Ostrowski said. "Our work with this grant will develop specific encoding strategies that exploit those strengths to represent large scenario spaces compactly."

One of the Ph.D. students involved will focus on the foundational work of developing and analyzing the quantum circuit encodings while the other works to evaluate how well the quantum method performs against classical benchmarks.

"UT's land-grant mission is about creating knowledge that serves the public, which also requires developing the next generation of researchers," said Ostrowski. "Quantum computing and operations research are both evolving rapidly, and training students at their intersection prepares a workforce that industry, national laboratories, and academia urgently need."

Open Source Results for Broader Impact

At the conclusion of the grant, Ostrowski and Herrman will not only publish their results but release multiple open source software libraries—including circuit templates, benchmark problems, simulation interfaces, and tutorials—designed for use by practitioners without expertise in quantum computing.

"Results in a paper tell you what happened; an open source code base lets others reproduce, extend, and build on the work," Ostrowski said. "It also lowers the barrier to entry for applied researchers and practitioners who want to engage with quantum optimization but don't have the background to build these tools themselves. Open source release is how research investments become community assets."

When the tools created during the grant become publicly available, researchers will be able to use the benchmarks as a common reference point to compare quantum and classical algorithms in future studies.

Meanwhile, industry practitioners in energy, logistics, and other sectors will have an accessible starting point for exploring whether quantum-enhanced optimization could improve their own decision-making workflows, Ostrowski explained.

"Ultimately, better stochastic optimization tools have downstream benefits for energy resilience, supply chain reliability, and emergency preparedness—areas that directly affect everyday citizens," he said. "Projects like this one are how a public university like UT translates federal investment in science into long-term human capital for Tennessee and the nation."

Provided by University of Tennessee at Knoxville