Researchers from IonQ and Oak Ridge National Laboratory have developed a new hybrid algorithm that uses both quantum and classical computing to solve a major problem in power grid operations: deciding which power generators to run and how much electricity each should produce.
The study, published on arXiv, focuses on the “unit commitment” (UC) problem—a complex task that power grid operators face every day. It involves selecting the best combination of generators to meet electricity demand while keeping costs low. As the number of generators and time intervals increases, the number of possible configurations becomes astronomically large.
By combining IonQ’s quantum hardware with classical optimization techniques, the researchers were able to generate cost-effective solutions for a variety of power grid scenarios. This included a realistic case with 26 generators and 24 hours of demand data.
The hybrid method could also be useful for other large-scale problems, such as airline crew scheduling and drug discovery. In such cases, the algorithm can quickly search through many options and find solutions close to the best possible one.
A Costly and Complex Problem
The unit commitment problem is crucial because it directly impacts electricity prices and energy reliability. Power grid operators must decide every hour which generators to activate and how much power each should supply. Though it sounds straightforward, the number of possible on-off combinations grows exponentially.
For instance, scheduling 15 generators over a day involves more than 10^108 possible configurations. Real-world constraints make the problem even harder: generators have power limits, switching them on and off incurs costs, and demand can change unexpectedly.
These challenges make the problem extremely difficult for traditional computers to solve in a reasonable time—especially for large power systems.
“Quantum computing is an emerging technology that can handle these types of complex optimization problems,” the researchers wrote. The U.S. Department of Energy’s 2024 roadmap even highlights power grid resource planning as a key use case for quantum-based methods.
A Three-Step Hybrid Approach
To address the problem, the team designed a three-step solution. First, they used a quantum algorithm to search for promising generator combinations. Next, they applied classical optimization to fine-tune the power output for each option. Finally, they selected the most cost-effective solution.
The quantum part used a technique called variational quantum algorithms (VQA), which adjust a quantum circuit to find the lowest “energy” state of a cost function. This cost function represents the generator schedule. To make the problem suitable for quantum hardware, the researchers translated it into a format called QUBO (Quadratic Unconstrained Binary Optimization).
They tested the algorithm using simulations and real runs on IonQ’s Forte quantum processor, which supports up to 36 algorithmic qubits. The tests involved grid models with 3, 10, and 26 generators across a full day of hourly time slots.
Strong Results with Small Errors
The hybrid algorithm produced results that were very close to optimal. In simulations, solutions were within 0.55% to 2.7% of the best possible outcome. For smaller systems, the algorithm found exact solutions. Even for the largest model, with 26 generators, the average error stayed around 2.5%.
When using IonQ’s actual quantum hardware, results closely matched the simulations. This suggests the algorithm can work well even on noisy quantum machines.
Importantly, the algorithm avoided the need to evaluate every possible configuration. Instead, it looked at only 128 carefully selected options per time step—a tiny slice of the total space.
Tradeoffs in Circuit Complexity
The study also explored how the depth of the quantum circuit affected results. Deeper circuits provided better accuracy but required more time and computing power. As the circuit’s complexity increased, so did the number of optimization steps.
This makes circuit design an important factor. The team used a smart layout that balanced complexity and performance. IonQ’s hardware, which allows full connectivity between qubits, helped reduce the number of required operations.
Looking Ahead
The researchers plan to improve the quantum part of the algorithm by using more advanced methods like variational quantum imaginary time evolution (varQITE). This approach lets the quantum system gradually settle into an optimal solution and may be better suited for constrained optimization problems like UC.
They also suggest designing custom quantum circuits that match the structure of power grid models. These changes could help scale the algorithm to real-world systems with hundreds of generators and more detailed constraints.
The project is part of a broader effort to apply quantum computing to energy systems. More efficient scheduling can lower costs, stabilize the grid, and reduce carbon emissions.
By showing that quantum computers—used alongside classical ones—can meaningfully support energy planning, the study adds momentum to the idea that quantum advantage may first appear in industrial optimization tasks, rather than in cryptography or chemistry.
