Researchers at the Department of Energy's (DOE) Oak Ridge National Laboratory (ORNL) have developed a computational method that could help electric grid operators and data center developers manage power more efficiently as artificial intelligence drives unprecedented growth in electricity demand.
The research, which was presented at the 2026 IEEE Power & Energy Society General Meeting, addresses a growing challenge as large data centers increasingly supplement utility power with their own on-site generation. Managing these additional generators adds significant complexity to power system operations, stretching the limits of today's grid optimization tools.
Researchers at ORNL developed a new "relax-and-round" optimization method that dramatically reduces the time needed to determine which generators should operate at any given time. The approach works by first "relaxing" the problem, temporarily allowing decisions that are normally restricted to yes-or-no choices to take on continuous values, so it can be solved much more quickly using efficient mathematical techniques. Once a solution is found, the method then "rounds" those values back to practical, discrete decisions about which generators should be on or off. This two-step process avoids the heavy computational burden of traditional mathematical optimization techniques while still producing high-quality, reliable results.
The team tested the approach using a system of 46 generators with different operating characteristics to represent a range of generator types. Researchers then created slightly modified copies of those generators to simulate variations within each type, allowing them to progressively scale the test systems in increments of 46 generators.
On a system with 276 generators, the new method cut computation time from about 10 hours to less than one second, with no increased cost or loss of feasibility. Larger systems became impractical to solve using the traditional method, while the new approach scaled efficiently to systems with tens of thousands of generators. That scalability could make the method a practical tool for planning future power grids as they accommodate large AI data centers and other distributed energy resources.
"The scientific community has developed powerful solvers for continuous variable problems, but not an efficient way to directly solve integer variable problems," said ORNL research scientist Shaked Regev, who led the project. "Instead of an all-purpose solver, we focused our attention on the specific problem of unit commitment and used powerful, problem-specific heuristics to reduce the integer variable problem to solving multiple continuous variable problems, which is faster and scales better."
Industry analysts project that U.S. data centers will add tens of gigawatts of new electricity demand over the next several years. As utilities work to expand grid capacity, many developers are investing in natural gas generators and other on-site power resources to ensure reliable operations.
Those new generating assets are expected to be smaller, more flexible and faster to respond than traditional power plants, requiring new approaches to scheduling and dispatch. ORNL's optimization method enables operators to quickly identify cost-effective operating plans, helping reduce fuel costs while maintaining reliability.
The algorithm is also designed for parallel computing and is compatible with graphics processing units (GPUs) opening the door to even faster performance on leadership-class computing systems.
The work highlights ORNL's expertise in power systems, applied mathematics and high-performance computing and supports DOE's efforts to modernize the nation's energy infrastructure while enabling continued growth in AI and advanced computing.
"Working in a multi-disciplinary team allows us to solve more complicated problems," said Regev. "I enjoyed learning a lot about power systems and teaching others about the quirks of mathematical optimization solvers. We were able to challenge assumptions about the correct way of doing things and this led us to the breakthrough."
The research was supported by ORNL's Laboratory Directed Research and Development program.
UT-Battelle manages Oak Ridge National Laboratory for DOE's Office of Science, the single largest supporter of basic research in the physical sciences in the United States. The Office of Science is working to address some of the most pressing challenges of our time. For more information, visit energy.gov/science . - Mark Alewine
