{"id":"beb74cc5-7583-44db-99cb-0f80036d0a9a","arxiv_id":"2505.18711","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Elastic wave equations are rewritten as Schrödinger-type systems and simulated on quantum hardware in principle, with gate-complexity bounds showing exponential advantage in dimension under periodic boundaries.","lead":"This paper applies a known quantum-simulation technique, Schrödingerisation, to elastic wave equations and derives algorithms with complexity bounds claiming exponential speedup in the spatial dimension. A smart generalist should read it because quantum simulation of wave propagation could eventually matter for seismology and materials science.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"p-domain scaling for p* is not tracked: recovery point grows as ε^{-1/r} or ε^{-1/2}, forcing N to grow with L and invalidating the stated gate counts as written.","rationale":"The reader's weakest assumption identifies the same load-bearing issue: the complexity proofs assume a fixed p-domain with N = O(ε^{-1}), while Theorem 5.3 requires p* to grow as ε^{-1/r} or ε^{-1/2}. My reading confirms and sharpens this: the p-domain half-width L must grow with p*, and once L grows, the Fourier resolution requirement changes from N = O(ε^{-1}) to N = O(ε^{-1} L), and the norm estimate ||H_s||max must be recomputed accordingly. This affects not only Theorems 5.1 and 5.2 but also Theorem 4.2 for the staggered-grid velocity-stress equations. The good news for the central claim is that the leading ε-exponents in the stated gate complexities survive after the correction: Theorem 5.1 remains ε^{-(1+1/r)}T and Theorems 4.2 and 5.2 remain ε^{-3/2}T, because the growth in N is exactly offset by the reduction in ||D_p||max. What changes is the qubit count and the log(1/ε) factor in the gate count, which must include log N. Thus the exponential advantage in spatial dimension d is plausible, but the theorems as written are not rigorously established; a conditional verdict with requested revisions is appropriate. I do not see a fatal flaw that would require rejection.","tokens_in":20771,"tokens_out":26168,"duration_ms":235151,"concrete_test":"Re-derive Theorems 4.2, 5.1, and 5.2 tracking the p-domain half-width L explicitly. Set L = c p* with p* from Eq. (3.10) and Theorem 5.3 (p*_s = O(ε^{-1/r}), p*_c = O(ε^{-1/2})), and set N = O(ε^{-1} L) so the spectral interpolation error of g(p) = e^{-p} on [-πL, πL] is at most ε. Verify whether the resulting gate count is O((⌈log dim⌉ + (d/r + 1 + 1/r) log(1/ε)) ε^{-(1+1/r)} T) for Theorem 5.1 and O((⌈log dim⌉ + (d/2 + 3/2) log(1/ε)) ε^{-3/2} T) for Theorems 4.2 and 5.2. If the omitted log N is required, the theorem statements as printed are incomplete; if N = O(ε^{-1}) is kept, show explicitly that p* lies outside the p-domain for small ε.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2 fixes recovery at p* ≥ max λ_n(H1) T (Eq. 3.10). For the staggered-grid and displacement Hamiltonians, ||H1||max = O(M), so Theorem 5.3 gives p*_s = O(ε^{-1/r}) and p*_c = O(ε^{-1/2}); Theorem 4.2 has the same issue with p* = O(ε^{-1/2}). Hence the p-domain half-width L must grow as O(p*) to contain the shifted wave packet. But the complexity proofs set N = O(ε^{-1}) and implicitly L = O(1), then bound ||H_s||max via ||D_p||max = O(N). Once L grows, the Fourier multiplier D_p has entries (l - N/2)/L, so ||D_p||max = N/(2L); to keep the spectral interpolation error of the initial profile e^{-p} at O(ε) over an interval of half-width L, the cutoff frequency N/L must be O(ε^{-1}), giving N = O(ε^{-1} L). This changes the proof bookkeeping: for Theorem 5.1, N = O(ε^{-(1+1/r)}) and ||H_s||max = O(ε^{-(1+1/r)}); for Theorems 4.2 and 5.2, N = O(ε^{-3/2}) and ||H_s||max = O(ε^{-3/2}). The final ε-exponents in Theorems 4.2, 5.1, and 5.2 happen to coincide with the stated ones, but the qubit counts are underestimated: the log(1/ε) term must include log N, i.e., an additional (1+1/r) or 3/2 factor, not just d/r or d/2. As written, the proofs contain an implicit L = O(1) that contradicts the recovery condition, so the central complexity theorems are not established exactly as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops quantum simulation algorithms for elastic wave equations using the Schrödingerisation framework of Jin, Liu, and Yu. It treats two formulations: the symmetric matrix form (SMF) of the velocity–stress equations in homogeneous media with external sources, using Fourier spectral discretization in space; and variable-coefficient problems, using a staggered-grid discretization of the velocity–stress equations and spectral or central-difference discretizations of a first-order hyperbolic system for the displacement. The main results are Hamiltonian simulation gate-complexity bounds (Theorems 4.1, 4.2, 5.1, 5.2) that scale polynomially in 1/ε and T and logarithmically in the spatial dimension d, and a recovery-point estimate (Theorem 5.3) for the auxiliary variable p. Classical numerical experiments in one and two dimensions compare the Schrödingerised solutions with standard finite-difference and spectral solutions and show good agreement.","tokens_in":21201,"tokens_out":20569,"duration_ms":161769,"significance":"Schrödingerisation-based quantum algorithms for elastic waves are a natural and potentially important application area, and the paper is among the first to give explicit complexity theorems for both velocity–stress and displacement formulations, including source terms and variable coefficients. The numerical verification of the recovery procedure is a clear strength, and the structural derivations (warped-phase transformation, homogenization of sources, Hamiltonian splitting) follow the established framework. If the complexity theorems can be repaired, the claimed exponential advantage in dimension for fixed final time is significant for geophysical and materials-science applications. However, the proofs currently contain a load-bearing gap in the treatment of the auxiliary p-domain, and the central-difference discretization as written is not a consistent first-derivative approximation; these issues must be resolved before the complexity claims can be accepted.","major_comments":[{"comment":"The recovery condition (3.10) requires p* ≥ max_n λ_n(H1) T, and Theorem 5.3 gives p*_s = O(ε^{-1/r}) and p*_c = O(ε^{-1/2}) because ||H1||max = O(M). The complexity proofs instead set N = O(ε^{-1}) (Eqs. (4.27) and (5.14)) and estimate ||H_s||max using ||D_p||max = O(N), which presupposes the p-domain half-length L is O(1). Once L must grow to O(p*) to contain the wave packet, the Fourier discretization of the initial profile e^{-|p|} requires the cutoff N/L = O(ε^{-1}), i.e., N = O(ε^{-1} L), and ||D_p||max = N/(2L) = O(ε^{-1}). This gives N = O(ε^{-(1+1/r)}) in Theorem 5.1 and N = O(ε^{-3/2}) in Theorems 4.2 and 5.2. The ε-exponents in the query-complexity factors are unchanged, but the qubit counts m_H and hence the stated gate complexities are missing the corresponding log(1/ε) terms, and the proof of the norm bound in Theorem 4.2 is not valid as written. The bookkeeping needs to be redone with L and N coupled to p*.","section":"§3.2, Theorems 4.2, 5.1, 5.2"},{"comment":"The matrix D defined in (5.15) is real symmetric (for periodic indexing it is essentially S + S^T) and does not approximate the first-derivative operator ∂_x appearing in the hyperbolic system (2.17). A standard second-order central difference for -i∂_x with periodic boundary conditions has the form (1/(2h))(-i)(S - S^T), with eigenvalues (1/h) sin(2πk/M); the matrix in (5.15) has eigenvalues 2 cos(2πk/M) and, when substituted into L_c in (5.16), makes the discretized operator skew-Hermitian in the constant-coefficient case, which is inconsistent with the H1/H2 split in (5.19) and with the eigenvalue estimate (5.26). The central-difference discretization, Theorem 5.2, and the p*_c estimate in Theorem 5.3 therefore need to be re-derived with the correct difference operator, or the definition of D needs to be clarified and justified.","section":"§5.2, Eq. (5.15)"}],"minor_comments":[{"comment":"The abstract and Section 3 title contain typos: 'expore', 'demontrate', 'dimensin', 'methed', and 'wraped'; these should be corrected.","section":"Abstract and Section 3 title"},{"comment":"The symbol p is used both for the pressure variable in (2.14) and for the auxiliary Schrödingerisation variable in Section 3; this is confusing and should be disambiguated.","section":"§2.1.2 and §3"},{"comment":"Remark 4.5 refers to 'Eqs. (2.17)' when discussing the classical complexity of the staggered-grid velocity–stress equations; it should refer to (2.6)/(4.20).","section":"Remark 4.5"},{"comment":"In the proofs of Theorems 4.1 and 5.1, the displayed qubit count contains an additional +log(1/ε) term that is dropped in the theorem statement; the two forms should be made consistent.","section":"Theorems 4.1 and 5.1"},{"comment":"The estimates (5.22)–(5.23) omit the factor T (or should state the assumption T = O(1)); also, 'Hermite matrix' in the proof should be 'Hermitian matrix'.","section":"Theorem 5.3"},{"comment":"The sparsity counts s = 3 and s = 6 appear to count only the spatial operator and not the tensor-product structure with D_p; since only constant factors are affected, this should be clarified.","section":"Theorems 4.1 and 4.2"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the paper is within scope for a quantum-computation or numerical-analysis journal. The main risk is not novelty but rigor: the p-domain scaling gap is likely fixable by reworking the proofs, and the central-difference definition appears to be a typographical error that can be corrected. I would encourage a major revision rather than rejection, provided the authors can supply a corrected complexity analysis and clarify the difference operator."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is the first Schrödingerised treatment of the symmetric-matrix velocity-stress system, the variable-coefficient staggered-grid form, and the displacement hyperbolic system, with explicit Hamiltonians and gate counts. The derivations are coherent, the numerical experiments reproduce the classical schemes, and the authors are honest about the periodic-boundary restriction. If the complexity theorems were exactly right, this would be a clean exponential-in-dimension speedup for elastic wave simulation.\n\nThe soft spots are in the complexity accounting, not in the basic method. The proofs of Theorems 4.2, 5.1, and 5.2 set N = O(ε^{-1}) for the p-direction while implicitly keeping the p-domain half-width L = O(1). But the recovery condition (3.10) and Theorem 5.3 force p* = O(ε^{-1/r}) or O(ε^{-1/2}), so L must grow. Then the Fourier cutoff N/L must be O(ε^{-1}) to resolve the initial profile, giving N = O(ε^{-1} L), and the qubit counts in the theorems are off: log N contributes (1+1/r) or 3/2 log(1/ε), not just the d/r or d/2 appearing in the displayed statements. The ε-exponents in the gate counts coincidentally survive, because ||H_s||_max = ||H1|| · ||D_p||_max ends up as O(M ε^{-1}) either way, but the proofs as written contain a contradictory L = O(1) assumption.\n\nThere is a second issue I rank as more serious. When p* grows, the recovered amplitude v(p*) = e^{-p*} u, so the probability of measuring the p-register in the recovery region is e^{-2p*}, super-polynomially small in ε. The theorems count Hamiltonian simulation queries only; they do not include the amplitude amplification (or huge sampling) needed to extract u. Unless the authors can show a different readout strategy, the end-to-end complexity has an extra e^{O(ε^{-1/r})} factor. For fixed ε and large d that still leaves an exponential advantage in d, but it is a significant omission in the stated claims.\n\nNet: the paper deserves a serious referee. The method is sound and the new PDE-specific work is real, but the complexity theorems need to be reworked, and the recovery probability needs to be quantified. I would send it out with a request for major revision.","headline":"Schrödingerisation applied to elastic wave equations with sound PDE derivations, but the complexity theorems have a real p-domain scaling gap and omit an exponentially small recovery probability; worth refereeing with major revision.","tokens_in":21696,"tokens_out":16505,"would_cite":false,"duration_ms":144979,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper shows that Schrödingerisation maps both velocity-stress and displacement forms of the elastic wave equations into Hamiltonian systems, yielding quantum gate complexities that grow polynomially in the spatial dimension while…","keywords":["elastic wave equations","Schrödingerisation","quantum simulation","Hamiltonian simulation","velocity-stress equations","staggered grid","spectral method","gate complexity"],"falsifier":"For the spectral-discretized displacement equation, compute the maximum eigenvalue of $H_1$ and the norm of the Schrödingerised Hamiltonian $H_s$ with $M=O(\\epsilon^{-1/r})$ spatial grid points, the $p$-domain length set to the $p_*$ of Theorem 5.3, and $N=O(\\epsilon^{-1})$ Fourier points in $p$; if $\\|H_s\\|_{\\max}$ grows faster than $O(\\epsilon^{-1})$, or the numerical error does not reach $O(\\epsilon)$, then the gate bound in Theorem 5.1 is incomplete.","tokens_in":20571,"feed_emoji":"🌊","tokens_out":10960,"duration_ms":73278,"temperature":0.7,"pith_summary":"The paper claims that the Schrödingerisation method—a warped phase transformation that embeds a linear PDE into one higher dimension and yields unitary evolution—can be applied to the elastic wave equations in two standard formulations, and that the resulting quantum algorithms have gate complexity scaling polynomially in the allowed error, the evolution time, and the logarithm of the spatial dimension. For the velocity-stress equations in isotropic media, the symmetric matrix form with spectral discretization and the variable-coefficient staggered-grid form both become Hamiltonian systems after Schrödingerisation, with gate counts of $O((\\log(d^2+3d)+(d/r)\\log(1/\\epsilon))\\epsilon^{-1}T)$ and $O((\\log(d^2+3d)+(d/2)\\log(1/\\epsilon))\\epsilon^{-3/2}T)$. For the displacement equation, the same treatment with spectral and central-difference discretizations gives analogous bounds, and a criterion (Theorem 5.3) selects the auxiliary $p$-domain size. Against the classical cost $O(\\epsilon^{-(d/r+1)}T^2)$, these bounds imply exponential quantum advantage in the spatial dimension $d$ whenever the solution is smooth and $T$ is $O(1)$. If the theorems hold, high-dimensional elastic wave simulation—relevant to seismic imaging and materials science—becomes a candidate for practical quantum advantage.","feed_headline":"Schrödingerisation gives exponential quantum speedup for elastic waves","feed_subtitle":"New gate bounds are polynomial in error, time, and log dimension, versus classical exponential in d.","key_machinery":"The central object is the Schrödingerisation (warped phase) transformation $v=e^{-|p|}u$, which maps a linear PDE $du/dt=(H_1+iH_2)u$ into $dv/dt=-H_1\\partial_p v+iH_2 v$, a one-dimension-higher Schrödinger-type equation. Discretizing $p$ by the Fourier spectral method and changing to Fourier coefficients gives a Hamiltonian system $d/dt\\,c=-i(H_1\\otimes D_p-H_2\\otimes I)c$, suitable for Hamiltonian simulation with quantum signal processing; the original solution is recovered at $p^*\\ge\\max\\{\\lambda_{\\max}(H_1)T,0\\}$ or by integration over $p$. The decomposition into Hermitian $H_1$ and anti-Hermitian $iH_2$ is what makes the transformation possible, and the sparsity of $H_1,H_2$ sets the gate complexity through the Low–Chuang query bound.","core_discovery":"The central discovery is that the non-unitary dynamics of elastic waves, including external forcing and spatially varying density and Lamé parameters, are exactly convertible into unitary Schrödinger-type evolution through the warped phase transformation $v=e^{-|p|}u$ and a discrete Fourier transform in the auxiliary variable $p$. This yields a Hamiltonian system $H=-i(H_1\\otimes D_p-H_2\\otimes I)$ whose sparsity is inherited from the original PDE, allowing Hamiltonian simulation via quantum signal processing. The paper proves gate-complexity bounds for four discretizations: the symmetric matrix form of the velocity-stress equation with spectral method (Theorem 4.1), the staggered-grid velocity-stress formulation for variable coefficients (Theorem 4.2), the displacement-equation hyperbolic system with spectral method (Theorem 5.1), and with central differences (Theorem 5.2). It also establishes, in Theorem 5.3, that the recovery point $p_*$ required to read out the solution grows as $O(\\epsilon^{-1/r})$ for the spectral method and $O(\\epsilon^{-1/2})$ for central differences, meaning the spectral method needs a larger auxiliary $p$-domain. The stated complexity is polynomial in $1/\\epsilon$, $T$, and $\\log d$, versus classical cost exponential in $d$, which the authors interpret as exponential quantum advantage in spatial dimension.","pith_inferences":["The exponential-in-$d$ advantage relies on $T$ being $O(1)$: for long-time simulations with $T\\sim\\epsilon^{-\\alpha}$, the quantum cost grows linearly in $T$ while the classical cost grows quadratically in $T$, so the advantage in dimension persists but the total speedup is moderated; the paper does not analyze this long-time regime.","Theorem 5.3's growth of $p_*$ may force the auxiliary $p$-domain length to grow with $1/\\epsilon$, which could invalidate the fixed $N=O(\\epsilon^{-1})$ Fourier lattice assumed in Theorems 5.1 and 5.2; a consistency check connecting $p_*$, domain length, and $N$ would settle whether the stated displacement-equation complexity is complete.","A natural testable extension is to apply the same Schrödingerisation construction to viscoelastic or anisotropic elastic wave equations, where the dispersion matrix $L(x,k)$ has different algebraic structure and the $H_1/H_2$ decomposition may change sparsity and the resulting gate counts.","The numerical experiments validate correctness on classical hardware but not the resource bounds; a resource estimate in logical qubits and T-gates for a concrete three-dimensional seismic problem would be needed to judge practical relevance."],"forward_implications":["For smooth solutions and evolution time $T\\sim O(1)$, the quantum gate count grows as $\\log d$ while classical cost grows as $\\epsilon^{-(d/r+1)}T^2$, giving exponential quantum advantage in spatial dimension $d$.","Variable medium parameters (position-dependent density and Lamé coefficients) are handled by the staggered-grid Schrödingerisation at cost $O(\\epsilon^{-3/2}T)$, enabling heterogeneous media simulation.","The displacement-equation formulation provides a practical rule for choosing the auxiliary $p$-domain: the recovery point scales as $\\epsilon^{-1/r}$ (spectral) or $\\epsilon^{-1/2}$ (central difference), and the numerical experiments confirm that both one-point and integral recovery match classical solutions.","External source terms in the symmetric matrix form are incorporated without spoiling unitarity via the homogeneous embedding $r=\\epsilon\\mathbf{1}$, so the Hamiltonian simulation framework extends to problems with forcing."],"supporting_citations":[{"why":"Introduces the Schrödingerisation method that maps general linear PDEs to Hamiltonian systems, the foundation of this paper's approach.","marker":"[JLY23]"},{"why":"Gives the optimal Hamiltonian simulation query complexity by quantum signal processing used in all gate-complexity theorems.","marker":"[LC17]"},{"why":"Establishes the staggered-grid method for the velocity-stress equations, the classical discretization adopted in Section 4.2.","marker":"[Vir86]"},{"why":"Develops the velocity-stress formulation with staggered grids and absorbing boundary conditions, supporting the variable-coefficient simulation.","marker":"[Ran89]"},{"why":"Provides the hyperbolic system form of elastic wave equations used in Section 5.","marker":"[RPK96]"},{"why":"Contributes the dispersion-matrix formulation and Hamiltonian-preserving scheme for high-frequency elastic waves that underlies Eq. (2.17).","marker":"[JL06]"},{"why":"Supplies the homogeneous-embedding technique for incorporating inhomogeneous source terms into Schrödingerisation.","marker":"[JLM24a]"}],"fun_headline_variants":["Quantum elastic wave solver: exponential advantage in spatial dimension","Elastic waves go quantum: Schrödingerisation enables exponential speedup","From PDE to unitary: quantum spectral method for elastic waves","Quantum simulation of elastic waves: dimension beats classical","Schrödingerisation turns elastic wave PDEs into quantum circuits"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The complexity proofs for the displacement equation assume a fixed auxiliary $p$-domain with $N=O(1/\\epsilon)$ Fourier points, while Theorem 5.3 requires the recovery point $p_*$ to grow as $(1/\\epsilon)^{1/r}$ or $(1/\\epsilon)^{1/2}$; if the $p$-domain must expand to contain $p_*$, the resolution count and the norm estimate $\\|H_s\\|_{\\max}=O(1/\\epsilon)$ used in Theorems 5.1 and 5.2 need reworking.","fun_headline_variants_meta":{"raw":{"variants":["Quantum elastic wave solver: exponential advantage in spatial dimension","Elastic waves go quantum: Schrödingerisation enables exponential speedup","From PDE to unitary: quantum spectral method for elastic waves","Quantum simulation of elastic waves: dimension beats classical","Schrödingerisation turns elastic wave PDEs into quantum circuits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000845,"raw_usage":{"total_tokens":3717,"prompt_tokens":1020,"completion_tokens":2697,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":636,"completion_tokens_details":{"reasoning_tokens":2616}},"tokens_in":636,"tokens_out":2697,"duration_ms":20270,"temperature":1.0,"reasoning_tokens":2616,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:28:10.719756+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the spectral-discretized displacement equation, compute the maximum eigenvalue of $H_1$ and the norm of the Schrödingerised Hamiltonian $H_s$ with $M=O(\\epsilon^{-1/r})$ spatial grid points, the $p$-domain length set to the $p_*$ of Theorem 5.3, and $N=O(\\epsilon^{-1})$ Fourier points in $p$; if $\\|H_s\\|_{\\max}$ grows faster than $O(\\epsilon^{-1})$, or the numerical error does not reach $O(\\epsilon)$, then the gate bound in Theorem 5.1 is incomplete.","supporting_citations":[],"review_version":1}