The Signal
Quantum Art Maps a Resonant Ring to a Wave-Simulation Algorithm
Quantum Art researchers propose a quantum algorithm for frequency-domain wave modeling that prepares only physically allowed Fourier modes. Their paper says the approach avoids a domain-size penalty in post-selection, while still requiring
Quantum.world Team · Written with AI · fact-checked ·
Quantum Art’s proposed wave-simulation algorithm takes aim at a conflict built into conventional quantum solvers: they can encode enormous grids compactly, but the mathematical systems used to solve wave problems can become harder to invert as the modeled domain grows. A second obstacle appears in the Fourier-space solution itself: physically relevant waves cluster near a resonant ring, so preparing a state across all possible modes and hoping to select that ring becomes increasingly inefficient. The company’s researchers propose preparing the ring directly instead.
The work appears in “Solving wave propagation problems via geometric quantum state preparation on dispersion manifolds,” posted to arXiv on September 9, 2026. A release carried by HPCwire on October 26 highlights the research. The paper presents an algorithm for frequency-domain wave propagation, not a report of the method running on a quantum processor.
The bottleneck sits on the resonance ring
The researchers formulate wave propagation as the Helmholtz equation, which describes steady-state fields at a given frequency. In the Fourier basis, the equation’s operator becomes diagonal, making a direct inversion look attractive. But the inverse becomes singular at resonance, where the squared wave number matches the squared frequency divided by the wave speed. In two dimensions, those resonant modes form a circle in Fourier space; in three dimensions, they form a sphere.
A generic quantum linear-systems solver pays for poor conditioning, which the paper says can grow with grid size. A direct Fourier-space arithmetic method faces a different cost: if it prepares all modes and post-selects only those near resonance, the useful fraction shrinks as the domain grows. The authors’ alternative constructs a quantum state on the resonant manifold first, then encodes source positions as phases on those modes.
“We address this challenge by directly preparing quantum states supported on the resonant manifold and encoding source locations through Fourier phases.”
Yakov Solomons, Lee Peleg, Netanel Barel, Jonathan Nemirovsky, Amit Ben-Kish and Yotam Shapira, Quantum Art
Domain size is not the only cost
The paper estimates gate complexity of O(n²), where n is the number of bits used to represent a spatial coordinate; that is polylogarithmic in the number of grid points. For a three-dimensional field spanning tens of kilometers at centimeter-scale wavelengths, the authors estimate a direct grid could require 10¹⁸ to 10²¹ points, compared with roughly 100 logical qubits for their quantum formulation. Those are scaling estimates, not a demonstration that a processor can run the full calculation at that scale.
Post-selection also remains. In the paper’s example, with a grid spacing of one-quarter wavelength and a resonant ring spanning nine Fourier-grid spacings, preparing the resonant state succeeds with probability about 2.3 × 10⁻³—roughly 450 attempts for one source. The source register adds a further factor that makes overall success probability fall as the number of sources rises. The favorable domain-size scaling therefore comes with a condition: the number of sources must stay small compared with the simulated grid.
Boundary conditions and hardware still matter
The main construction assumes periodic boundary conditions, a natural fit for Fourier transforms. The authors say Dirichlet or Neumann boundaries can be handled with odd or even extensions, at the cost of one additional qubit per spatial dimension. They also propose extending the approach to certain anisotropic or elastic wave problems when the governing operator can be diagonalized in a spectral basis and the solution occupies a lower-dimensional manifold.
The paper’s evidence is an algorithmic analysis with simulated examples, not a hardware benchmark. It does not establish that the estimated logical-qubit counts translate directly into a practical system, or that preparing and using the encoded solution will beat classical methods for a complete application. For now, the concrete result is a circuit strategy: build the resonant modes directly, attach source-dependent phases, and avoid searching the full Fourier grid for a wave pattern the physics confines to a ring.