Benchmarking Portfolio Optimization with Quantum Annealing

Photograph of D-Wave processor


Researchers from the Oak Ridge National Laboratory (ORNL) Quantum Computing Institute benchmarked the performance of the D-Wave 2000Q quantum annealer for portfolio optimization problem. In particular, they demonstrated that tuning parameters could be implemented on the quantum annealer to improve the probability of success in fining the global minimum solution. Reverse annealing strategies were implemented and yielded an order of magnitude improvement in probability of success over the standard forward annealing controls. These results show that hardware controls can have a significant impact on performance for difficult optimization problems as seen in figure 1. It was also found that certain hardware controls had little to no effect on performance.

Significance and Impact

These results advance the understanding of how quantum computers may impact future DOE mission in resource optimization and allocation. By demonstrating how to solve problems in constrained combinatorial optimization, this work provides new tool for future applications in energy distribution and transmission networks.

Research Details

  • Sample portfolio optimization problems were generated using a uniform random distribution of stock prices for a set of assets. The number of assets in a portfolio included 2, 3, 4, and 5 while considering 1000 sample problems for each problem size.
  • Quantum annealing hardware controls were investigated including forward annealing time, reverse annealing, problem embedding, and spin reversal transforms.
  • The benchmarking techniques used included probability of success to find the global minimum solution for each problem and the probability of noise in the hardware causing chains to break and no longer fully represent a logical state.


Sponsor/Funding: DOE Early Career Research Program

PI and affiliation: Travis Humble, Quantum Computational Sciences, Quantum Computing Institute, Quantum Science Center

Team: Erica Grant (UTK/ORNL), Travis Humble (ORNL), Benjamin Stump (ORNL)

Citation and DOI: E. Grant, T. S. Humble, and B. Stump, “Benchmarking Quantum Annealing Controls with Portfolio Optimization,” Phys. Rev. Applied 15, 014012 (2021)

DOI: 10.1103/PhysRevApplied.15.014012

Last Updated: February 19, 2021 - 2:55 pm