Practical noise mitigation for quantum annealing via dynamical decoupling: Toward industry-relevant optimization using trapped ions

Sebastian Nagies, Chiara Capecci, Marcel Seelbach Benkner, Javed Akram, Sebastian Rubbert, Dimitrios Bantounas, Michael Moeller, M. Johanning, Philipp HaukePhysical 2026 · DOI

TL;DR

This paper presents a practical method to combat noise in quantum annealing, a promising technique for optimization problems. By applying "dynamical decoupling" pulses, the researchers show they can restore near-ideal performance even with significant magnetic field fluctuations, a major hurdle for current quantum hardware. This work offers a scalable solution for building more reliable quantum annealers for real-world applications.

What Problem Does This Paper Address?

Quantum annealing is a powerful approach for tackling complex optimization problems, like those found in logistics or machine learning. It works by guiding a quantum system towards its lowest energy state, which encodes the optimal solution. However, real-world quantum annealers are plagued by environmental noise, particularly fluctuations in magnetic fields. These fluctuations can corrupt the system's lowest energy state, leading to incorrect solutions and hindering the practical application of this technology. This paper aims to address this critical limitation by developing and demonstrating a noise mitigation strategy.

Methods

The core of the paper's solution lies in dynamical decoupling (DD). This technique involves applying precisely timed control pulses to the quantum bits (qubits) to effectively "cancel out" the effects of environmental noise. Specifically, the authors use global spin-flip pulses. They also adapt the standard quantum annealing protocol by keeping the "cost Hamiltonian" (which encodes the problem) constant and gradually reducing the "driving Hamiltonian" (which introduces quantum fluctuations). This modified protocol, combined with DD pulses, is shown to be particularly well-suited for trapped-ion quantum computers using Magnetic Gradient Induced Coupling (MAGIC), a platform offering all-to-all qubit connectivity. The researchers also explore an alternative encoding scheme that isolates local field noise, making it easier to mitigate with DD.

Datasets

The paper uses several benchmark optimization problems to test their approach:

  • Minimal Multiple Object Tracking (MOT) QUBO instances: These are small, five- and nine-qubit problems derived from computer vision tasks, designed to be representative of real-world challenges.
  • Cutting Stock instances: Five- and six-qubit problems based on a classic logistics optimization challenge.
  • Sherrington-Kirkpatrick (SK) model: A standard benchmark for spin-glass problems, used to test the protocol's robustness across different problem structures and sizes (up to 12 qubits).

The authors do not introduce new datasets but rather use these established problem formulations to evaluate their noise mitigation strategy.

Key Insights and Ideas

  • Local field noise is the primary culprit: While fluctuations in two-body couplings have a negligible impact on performance in realistic scenarios, local magnetic field fluctuations significantly degrade annealing fidelity.
  • Dynamical decoupling is highly effective: Even moderate rates of DD pulses (achievable with current technology) can restore annealing fidelity to near-ideal levels, even for substantial noise amplitudes.
  • A universal scaling law emerges: The fidelity of the annealing process, when affected by local field noise and mitigated by DD, can be described by a generalized parameter that combines noise amplitude and the time between DD pulses. This suggests a fundamental principle governing noise mitigation.
  • Trapped-ion platforms are well-suited: The MAGIC architecture in trapped-ion systems, with its native all-to-all connectivity and tunable couplings, is presented as a promising platform for implementing these noise-mitigated quantum annealing protocols.
  • Scalability is demonstrated: The protocol's effectiveness is shown to hold across different problem sizes and types, indicating its potential for industry-relevant applications.

Key Results

  • Significant fidelity restoration: The proposed DD protocol can restore annealing fidelity to levels close to the noiseless case, even with noise amplitudes up to 1000 Hz (significantly larger than the characteristic energy scale of the annealing Hamiltonian).
  • Achievable pulse rates: Approximately 2.5 dynamical decoupling pulses per millisecond are sufficient to achieve good performance, a rate well within the capabilities of current trapped-ion systems.
  • Robustness across problem types: The effectiveness of the DD protocol is demonstrated across MOT, cutting stock, and SK model problems, showing little dependence on the specific problem structure or size.
  • Universal scaling confirmed: Analytical and numerical results show that fidelity loss is governed by the product of noise amplitude and the time between DD pulses, with a slight modification for time-dependent noise.

Limitations and Future Work

  • Analytical derivation of generalized parameter: The authors suggest that a deeper analytical understanding of the generalized scaling parameter and its dependence on noise spectral density would be valuable.
  • Additional noise sources: Future work could investigate the impact of other noise sources and explore more sophisticated dynamical decoupling protocols.
  • Broader range of optimization problems: Applying the strategy to a wider variety of optimization problems would further validate its universality.
  • Quantum optimization algorithms: Incorporating DD sequences into quantum optimization algorithms could allow for controlled studies of noise and quantum effects.
  • Near-term experimental demonstrations: The work is expected to motivate experimental demonstrations on near-term quantum annealers and inform error mitigation strategies for other architectures.

Summary updated

Abstract

Quantum annealing is a framework for solving combinatorial optimization problems. While it offers a promising path toward a practical application of quantum hardware, its performance in real-world devices is severely limited by environmental noise that can degrade solution quality. We investigate the suppression of local field noise in quantum annealing protocols through the periodic application of dynamical decoupling pulses implementing global spin flips. As test problems, we construct minimal multiple-object-tracking quadratic unconstrained binary optimization instances requiring only five and nine qubits, as well as cutting-stock instances of five and six qubits. Moreover, using the Sherrington-Kirkpatrick model, we demonstrate the robustness of our protocol to problem structure and size. To further place our results in a practical context, we consider a trapped-ion platform based on magnetic-gradient-induced coupling as a reference architecture, using it to define experimentally realistic noise and coupling parameters. We show that external magnetic field fluctuations, typical in such setups, significantly degrade annealing fidelity, while moderate dynamical decoupling pulse rates, which are achievable in current experiments, restore performance to near-ideal levels. Our analytical and numerical results reveal a universal scaling behavior, with fidelity determined by a generalized parameter combining noise amplitude and dynamical decoupling pulse interval. While our analysis is grounded in the trapped-ion platform, the proposed noise-mitigation strategy and resulting performance improvements are applicable to a broad range of quantum annealing implementations and establish a practical and scalable route for error mitigation in near-term devices.
BibTeX
@inproceedings{nagies2026practical,
  author = {Sebastian Nagies and Chiara Capecci and Marcel Seelbach Benkner and Javed Akram and Sebastian Rubbert and Dimitrios Bantounas and Michael Moeller and M. Johanning and Philipp Hauke},
  title = {{Practical noise mitigation for quantum annealing via dynamical decoupling: Toward industry-relevant optimization using trapped ions}},
  booktitle = {Physical},
  year = {2026},
  doi = {10.1103/mfqh-jtrl},
}
Figure