Rigetti has developed a quantum preconditioning algorithm that enhances optimization processes by accelerating classical solvers’ convergence on mathematical benchmarks and a real-world power grid problem using their 84-qubit Ankaa-3 system.
Tests showed significant speed improvements, with Burer-Monteiro achieving up to 9x faster convergence for maximum-cut problems and high accuracy in fewer iterations, even at low preconditioning levels.
Random 3-Regular Graph Maximum-Cut Problems
The work explores the application of quantum preconditioning to a power grid optimization problem using South Carolina’s energy grid data. The goal was to compute the maximum power exchange section, crucial for assessing grid health and power delivery capability. Quantum preconditioning was implemented on Rigetti’s 84-qubit Ankaa-3 computer, optimizing over 410 variables with the Burer-Monteiro solver, which is more efficient than simulated annealing.
Results showed that quantum preconditioning enabled the solver to achieve 99.95% accuracy in four times fewer iterations compared to the original problem. While classical methods can reach higher accuracy with more iterations, real-world constraints often limit time, making faster high-accuracy solutions valuable. The study highlights benefits even with a low preconditioning level (p=1) and imperfect fidelities.
Looking ahead, the document emphasizes considering preconditioning time when assessing advantages. Faster convergence of classical solvers post-preconditioning was observed, establishing thresholds where preconditioning is justified. Superconducting platforms like Ankaa-3 offer advantages due to fast execution times, with future focus on optimizing preconditioning time for quantum advantage in optimization.
Enhancing Power Grid Optimization
The application of quantum preconditioning to South Carolina’s energy grid optimization demonstrates its potential in enhancing computational efficiency and accuracy. By leveraging Rigetti’s Ankaa-3 quantum computer, researchers optimized over 410 variables using the Burer-Monteiro solver, achieving a remarkable 99.95% accuracy with only a quarter of the iterations required without preconditioning. This highlights the value of reducing computation time in real-world scenarios where resources are often constrained.
Despite operating under a low preconditioning level (p=1) and dealing with imperfect quantum gate fidelities, the approach still delivered significant benefits. The study underscores that even minor adjustments can yield substantial improvements, suggesting that quantum methods may not require perfect conditions to be effective.
The integration of quantum preconditioning with classical solvers shows promise for faster convergence after the initial quantum step. Superconducting platforms like Ankaa-3, capable of executing circuits in microseconds, are pivotal in minimizing preconditioning time. Continued optimization of this process could establish a clear quantum advantage in practical optimization tasks.
This work is part of Rigetti’s broader efforts in quantum optimization algorithms and has received support from the U.S. Department of Energy’s SQMS Center, focusing on advancing superconducting quantum technologies. The findings suggest that combining quantum techniques with classical methods can address real-world challenges efficiently, even with current hardware limitations.
More information
External Link: Click Here For More
