From Hope to Heuristic: Realistic Runtime Estimates for Quantum Optimisation in NHEP

\firstname Maja \lastname Franz \fnsep, \firstname Manuel \lastname Schönberger, \firstname Melvin \lastname Strobl \fnsep, \firstname Eileen \lastname Kühn, \firstname Achim \lastname Streit, \firstname Pía \lastname Zurita, \firstname Markus \lastname Diefenthaler, \firstname Wolfgang \lastname Ma · 2025

Read From Hope to Heuristic: Realistic Runtime Estimates for Quantum Optimisation in NHEP by \firstname Maja \lastname Franz \fnsep, \firstname Manuel \lastname Schönberger, \firstname Melvin \lastname Strobl \fnsep, \firstname Eileen \lastname Kühn, \firstname Achim \lastname Streit, \firstname Pía \lastname Zurita, \firstname Markus \lastname Diefenthaler, \firstname Wolfgang \lastname Ma (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-05066.txt.

Introduction

Maja Franz and seven co-authors price, in 2025 work, how long quantum optimizers would truly take on a hard track-find task from particle physics. Near-term quantum boxes are flawed, yet paired with fit drills they may pay. Two drill kinds are priced: heat-settle runs and a staged wave drill.

Track finds are cast in the square two-state form both drills eat. Low tones in the wave drill's knob space prove key to good settle plans, the team finds. That law seeds a shortcut rule that saves means while nearing the best end. Craft the drill with the box, they urge.

Catalog record

CREATOR\firstname Maja \lastname Franz \fnsep, \firstname Manuel \lastname Schönberger, \firstname Melvin \lastname Strobl \fnsep, \firstname Eileen \lastname Kühn, \firstname Achim \lastname Streit, \firstname Pía \lastname Zurita, \firstname Markus \lastname Diefenthaler, \firstname Wolfgang \lastname Ma
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-05066

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

From Hope to Heuristic: Realistic Runtime Estimates for Quantum Optimisation in NHEP \firstname Maja \lastname Franz \fnsep, \firstname Manuel \lastname Schönberger, \firstname Melvin \lastname Strobl \fnsep, \firstname Eileen \lastname Kühn, \firstname Achim \lastname Streit, \firstname Pía \lastname Zurita, \firstname Markus \lastname Diefenthaler, \firstname Wolfgang \lastname Mauerer Abstract nisq ( nisq ) computers, despite their limitations, present opportunities for near-term quantum advantages in nhep ( nhep ) when paired with specially designed quantum algorithms and processing units. This study focuses on core algorithms that solve optimisation problems through the quadratic Ising or quadratic unconstrained binary optimisation model, specifically quantum annealing and the qaoa ( qaoa ). In particular, we estimate runtimes and scalability for the task of particle tr ( tr ), a key computing challenge in nhep , and investigate how the classical parameter space in qaoa , along with techniques like a Fourier -analysis based heuristic, can facilitate future quantum advantages. The findings indicate that lower frequency components in the parameter space are crucial for effective annealing schedules, suggesting that heuristics can improve resource efficiency while achieving near-optimal results. Overall, the study highlights the potential of nisq computers in nhep and the significance of co-design approaches and heuristic techniques in overcoming challenges in quantum algorithms. QC Quantum Computing NISQ Noisy Intermediate-Scale Quantum NHEP Nuclear and High-Energy Physics QA Quantum Annealing QAOA Quantum Approximate Optimisation Algorithm QUBO Quadratic Unconstrained Binary Optimisation LHC Large Hadron Collider TR Track Reconstruction SG Spectral Gap RI Random Parameter Initialisation 1 Introduction nisq computers, while limited by imperfections and small scale, hold promise for near-term quantum advantages in nhep when coupled with co-designed quantum algorithms and special-purpose quantum processing units franz_co-design_2024 . Developing co-design approaches is essential for near-term usability, but inherent challenges exist due to the fundamental properties of nisq algorithms preskill_quantum_2018 . In this contribution we therefore investigate the core algorithms, which can solve optimisation problems via the abstraction layer of a quadratic Ising model or general qubo ( qubo ), namely qa ( qa ) and the qaoa . As an example for a variety of applications in nhep utilising qubo formulations pakhomchik_solving_2022 , we focus on tr okawa_quantum-annealing-inspired_2024 ; zlokapa_charged_2021 . tr is a good example of pattern recognition and optimisation in nhep , and with TrackML Kiehn_2019 , there exist curated datasets for algorithmic development along well-documented performance metrics. While qa and qaoa do not inherently imply quantum advantage, qa runtime for specific problems can be determined based on the physical properties of the underlying Hamiltonian, although it is a computationally hard problem itself albash_adiabatic_2018 . Our primary focus is on two key areas: Firstly, we estimate runtimes and scalability for the common nhep problem of tr addressed via a qubo formulation bapst_pattern_2019 . This analysis is conducted by identifying minimum energy solutions of intermediate Hamiltonian operators encountered during the annealing process. Secondly, we investigate how the classical parameter space in the qaoa , together with approximation techniques such as a Fourier -analysis based heuristic, proposed by Zhou et al. zhou_quantum_2020 , can help to achieve (future) quantum advantage, considering a trade-off between computational complexity and solution quality. The remainder of this article is structured as follows: In Sec. 2 we provide a brief introduction to qa and explain how annealing schedules and energy levels link to sg . Sec. 3 then introduces qaoa , how it can be used to derive annealing schedules, and what heuristics exist for doing so. In Sec. 4 , we briefly introduce the task of tr and describe our experimental setup with the results in Sec. 5 . Finally, we conclude in Sec. 6 . We ensure reproducibility mauerer:22:q-saner , by providing the code as well as the numerical results that we use in the figures of this work in Ref. franz_2025_14921650 . Notably, our experiments are performed using classical simulations of qaoa and qa with Qiskit qiskit2024 . 2 Quantum Annealing qa is, roughly speaking, a restricted version of adiabatic quantum computing albash_adiabatic_2018 which is used to find minimum energy solutions to a problem (cost) Hamiltonian \hat{H}_{C} . The solutions are obtained by an adiabatic transition from a system, prepared in the ground state of an initial Hamiltonian \hat{H}_{0} , to \hat{H}_{C} . qa is restricted to a certain subclass of all possible Hamiltonians albash_adiabatic_2018 , which are equivalent to qubo problems, to which any NP problem can be reduced lucas_ising_2014 . The overall time-dependent Hamiltonian of the annealing process is determined by the “protocol” f(t) that guides the transition between the Hamiltonians via \hat{H}_{QA}(t)=-\left(f(t)\hat{H}_{C}+(1-f(t))\hat{H}_{0}\right) (1) By default, D-Wave annealers dwave implement a linear transition given by f(t)=s\coloneqq t/T , where T is the chosen annealing time , and t\in[0,T] . It is known that the minimum time T_{\text{min}} required for an adiabatic transition depends on the spectrum of the Hamiltonian \hat{H}_{QA} , in particular the minimum energy gap ( sg \Delta_{\text{min}} ) between ground state and first excited state encountered when running the protocol given in Eq. 1 by T_{\text{min}}=\mathcal{O}(1/\Delta_{\text{min}}^{2}) . Harder problems with longer required runtimes therefore comprise smaller minimum sg . Figure 1 : Left: Energy values for ground and first excited state of the Hamiltonian given in Eq. 1 over interpolation parameter s for an example tr problem instance considered in this work. Right: Corresponding sg ( i.e. , the difference between energy levels of the first excited and the ground state). Lines do not provide an interpolation, but are only used to guide the eye; insets magnify the region around the minimal sg . Consider, as shown on the left-hand side of Fig. 1 , the s -dependent discrete energy levels of the ground state and the first excited state of one of the tr problem instances considered in this work (see Sec. 4 ). Physical intuition suggests that there are regions with a large gap between the two states, where the annealing process can proceed more rapidly without the encountering level transitions. Conversely, in regions characterised by a small sg , the annealing process must conducted at a slower rate to prevent the system from entering higher energy states, as this would result in sub-optimal or invalid solutions. The minimum sg is observed at around s=0.85 and persists at low levels until s=1 . The annealing schedule, denoted by f(t) , can be adapted to accommodate the properties of the problem under consideration. Intuitively, this means that during the first three-quarters of the time, the transition from Hamiltonian \hat{H}_{0} to \hat{H}_{C} can proceed swiftly, and that it needs to slow down in the last quarter. However, givent the computational complexity of the energy spectrum of the Hamiltonian ( Eq. 1 ) and its unavailability prior to the problem’s formulation, alternative means of determining good schedules are necessary. 3 Quantum Approximate Optimisation Algorithm One such possibility is to use insights gained from qaoa computations. As qa , the qaoa is designed to optimise qubo problems. It employs a quantum circuit with p\in\mathbb{N} layers of unitary operators defined by 2p parameters \vec{\beta},\vec{\gamma}\in\mathbb{R}^{p} . A qaoa layer j\in[1,p] comprises two unitaries: U_{M}(\beta_{j})=e^{-i\beta_{j}\hat{H}_{M}}\quad\mathrm{and}\quad U_{C}(\gamma_{j})=e^{-i\gamma_{j}\hat{H}_{C}} (2) with U_{M} representing mixer Hamiltonian \hat{H}_{M} , and U_{C} based on the cost Hamiltonian \hat{H}_{C} , of which the ground state encodes the optimal solution to

Read the complete text