A new method tracks how entanglement and coherence change during optimisation processes on near-term quantum computers. The investigation reveals distinct patterns in resource use within two leading algorithms, QAOA and FALQON, as they seek solutions. Researchers at the Federal University of Santa Maria discovered that QAOA and FALQON utilise entanglement and coherence differently when seeking solutions; these properties are fundamental to how quantum computers work.

QAOA builds up then reduces entanglement as it progresses, while FALQON exhibits lower initial levels followed by gradual decay alongside improved results. Understanding these distinct patterns in resource use is key for designing more effective quantum algorithms before advanced hardware becomes available. Researchers at the Federal University of Santa Maria shed light on how key quantum properties, entanglement and coherence, behave within optimisation algorithms designed for near-term quantum computers; these resources aren’t used equally across different approaches.

The team investigated QAOA and FALQON, revealing that QAOA initially builds up then reduces entanglement as it searches for solutions, while FALQON demonstrates lower initial levels followed by gradual decay alongside improved results. Understanding this ‘resource use fingerprint’, quantified using techniques like measuring distance between quantum states akin to comparing fingerprints, is vital for improving algorithm design before more powerful hardware arrives.

Uhlmann’s theorem allows researchers to assess entanglement without fully probing the internal state of qubits, much like diagnosing an illness from external symptoms than invasive tests, and was key to their analysis; but what do these differing resource trajectories mean for optimising performance on noisy intermediate-scale quantum devices.

QAOA and FALQON exhibit distinct Bures entanglement dynamics during optimisation

Federal University of Santa Maria researchers quantified Bures entanglement, a measure of connection between qubits, revealing that QAOA initially accumulates this property before reducing it, peaking prior to decline. While FALQON reached substantially lower peaks later in execution, Uhlmann’s theorem enabled indirect measurement via auxiliary circuits akin to diagnosing symptoms rather than invasive tests; this contrasts with previous methods requiring complete state tomography which proved impractical for larger systems.

These differing resource trajectories demonstrate each algorithm’s unique use of quantum resources and provide key insights into optimising performance on near-term devices where coherence is fragile and qubit counts are limited. Specifically, the Federal University of Santa Maria team quantified not only entanglement but also ‘Bures coherence’.

They found that QAOA’s coherence generally decreased with each added layer of computation, though some residual value remained due to solution ambiguity. In contrast, FALQON exhibited smoothly decreasing coherence alongside improving optimisation results, its coherence fell consistently as it neared the correct answer; auxiliary circuits achieved these new measurements by effectively diagnosing resource use without disrupting the main calculation, a significant improvement over previous methods needing complete state tomography impractical for systems beyond a few qubits.

Entanglement and Coherence Tradeoffs Inform Near Term Quantum Optimisation Performance

Building genuinely useful quantum computers requires understanding how algorithms like QAOA and FALQON utilise fundamental properties such as entanglement, that bizarre connection between qubits, and coherence, given their limited availability on near-term devices. The team acknowledge their analysis focused solely on one specific instance: finding the smallest set of vertices to cover all edges in a graph. Despite this focus on identifying the smallest connections within a network, this work from the Federal University of Santa Maria offers valuable insight into the behaviour of two leading optimisation algorithms at a fundamental level.

Determining whether Quantum Approximate Optimisation Algorithm (QAOA) or Feedback-based Algorithm for Quantum Optimisation (FALQON) rely more heavily on ‘quantumness’, specifically entangled qubits and their coherence, is important as hardware improves; essentially, it reveals how long they maintain information. The analysis revealed QAOA initially builds up then reduces entanglement while progressing towards solutions, retaining some residual coherence due to ambiguity in potential answers. Conversely, FALQON exhibits lower peak levels of qubit linkage alongside decreasing coherence correlated with improved performance on the Minimum Vertex Cover problem, offering detailed insights that will help refine these optimisation strategies as more powerful hardware emerges.

This research investigated the use of quantum resources, entanglement and coherence, within two variational algorithms, QAOA and FALQON, when solving a minimum vertex cover problem. Results showed QAOA accumulated entanglement early in its process before reducing it upon finding a solution, while maintaining some level of coherence; conversely, FALQON demonstrated consistently lower entanglement paired with smoothly declining coherence as optimization progressed.

These findings provide insight into how each algorithm utilizes quantum properties during computation without requiring complete state tomography, which is impractical for larger systems. This understanding informs refinement of these optimisation strategies as more powerful hardware becomes available.

👉 More information
🗞 Resource Analysis of Quantum Optimization Algorithms via Variational Methods
✍️ Vitor Dallabona, Lucas Friedrich and Jonas Maziero (Federal University of Santa Maria)
🧠 ArXiv: https://arxiv.org/abs/2610.29919

More like this