Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Uniform semiclassical observable error bound of Trotter-Suzuki splitting: a simple algebraic proof

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Trotter–Suzuki splitting computes polynomial observables of the semiclassical Schrödinger equation with error $O(\Delta t^p)$ independent of $h$, for arbitrary even order $p$.

desk verdict The local algebraic error analysis is solid, but the global Theorem 2.1 is unproven: the paper skips the telescoping argument that would require uniform control on Heisenberg-evolved observables. read the letter →

arxiv 2507.02783 v2 pith:MC3XTFAQ submitted 2025-07-03 math.NA cs.NAquant-ph

Di Fang , Conrad Qu This is my paper
classification math.NAcs.NAquant-ph MSC 35Q4165M1581Q2068Q12
keywords Trotter-SuzukisplittingsemiclassicalSchrödingerequationuniform-in-hobservableerrorcommutatorscalingheight-widthalgebrafinitedifferencediscretizationspectralquantumsimulation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that for the semiclassical Schrödinger equation, a $p$-th order Trotter–Suzuki splitting computes time-evolved polynomial observables with error $\|T_{p,n}(\Delta t) - T(t)\| \le C \Delta t^{p}$, where $C$ does not depend on the semiclassical parameter $h$ (Theorem 2.1). This matters because unitary or wavefunction simulation needs $\Delta t$ to shrink as $h \to 0$, so an $h$-independent step would remove a major cost from grid-based semiclassical simulation; the number of Trotter steps for precision $\epsilon$ becomes $O(\epsilon^{-1/p})$. The proof is algebraic: it introduces a height and width for differential operators and shows that every nested commutator entering the local error has discretized norm $O(1)$ uniformly in $h$, avoiding Egorov-type theorems and semiclassical limits. The paper also states a general nested-commutator observable bound valid for any bounded Hamiltonian $H = A + B$, beyond the semiclassical regime.

What carries the argument

The carrying object is the height-width structure of the Lie algebra generated by $A = -\frac{h}{2}\partial_x^2$ and $B = h^{-1}V$. For an operator written as a sum of terms $y(x)h^m\partial_x^d$, its height is the largest derivative order $d$ and its width is the smallest power $m$ of $h$. Lemma 4.1 states that commutators reduce height, $\operatorname{ht}([P,Q]) \le \operatorname{ht}(P)+\operatorname{ht}(Q)-1$, and expand width, $\operatorname{wd}([P,Q]) \ge \operatorname{wd}(P)+\operatorname{wd}(Q)$; the discrete analogue (Lemma 4.7) holds for finite-difference matrices. Since a derivative order $d$ costs $h^{-d}$ after spatial discretization while width $m$ contributes $h^{m}$, any nested commutator with an observable monomial $y_m(x)h^m\partial_x^m$ satisfying $\operatorname{ht} \le \operatorname{wd}$ has norm $O(1)$, uniformly in $h$. The Taylor expansion of the exact and Trotter-evolved observables reduces the local error to exactly such nested commutators, so the $h$-independence transfers from commutators to the error.

What would settle it

Pick $V(x)=\cos(x)$ on a periodic grid, observable $O = h\partial_x$ (or $O = h\partial_x + h^2\partial_x^2$), $p=2$, $t=1$, and $\Delta t = 0.1$. Compute the global error $\|T_{2,n}(\Delta t)-T(t)\|$ for $h = 1/32, 1/64, \dots, 1/1024$. If the error grows with $h^{-1}$, or grows with the number of steps $n$ at fixed $\Delta t$ and small $h$, the claimed uniform bound fails; the paper's Figure 3 shows flatness for its chosen setup, so a longer-time or higher-order-observable version of that plot directly probes the missing accumulation step.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a uniform-in-$h$ observable error bound for arbitrarily high-order Trotterization. For discrete operators $A = -(h/2)D^2$ and $B = h^{-1}\operatorname{diag}(V)$ from spectral or finite-difference discretizations, and observables $O = \sum_{m=0}^{q} y_m(x) h^m \partial_x^m$ with smooth bounded coefficients, the global Heisenberg-picture error satisfies $\|T_{p,n}(\Delta t) - T(t)\| \le C \Delta t^{p}$ with $C$ independent of $h^{-1}$ (Theorem 2.1), for $t = n\Delta t = O(1)$. The local one-step error is $O(\Delta t^{p+1})$ (Theorem 3.1), and the uniformity comes from a height-width lemma: a commutator lowers the highest derivative order by at least one and raises the smallest $h$-power, so any $(p+1)$-fold commutator with such an observable has height no greater than width, and after discretization its norm is $O(1)$ as $h \to 0$. This gives the paper's claimed first purely algebraic proof of uniform observable bounds for high-order Trotter–Suzuki formulas, with no Egorov or semiclassical-limit ingredients.

Load-bearing premise

The load-bearing premise is that the error bound proved for a single time step continues to hold when the errors from all $n$ steps are added up, with $h$ never creeping back into the constant; the paper states this but does not supply the invariance argument that would justify the summation.

Editorial extensions

If this is right

  • For any even order $p$, achieving observable accuracy $\epsilon$ needs $n_p = O(\epsilon^{-1/p})$ Trotter steps, with no factor depending on $h$ (Corollary 2.3).
  • Both spectral and finite-difference spatial discretizations are covered, so the $h$-independent step applies to practical grid-based semiclassical simulation of polynomial observables.
  • The local error is $O(\Delta t^{p+1})$ with an $h$-independent constant; the paper's Theorem 2.1 upgrades this to a global $O(\Delta t^{p})$ bound for $t = O(1)$ after $n = t/\Delta t$ steps.
  • Proposition 2.5 gives an operator-norm observable error bound for any bounded $H = A + B$ in terms of $(p+1)$-fold nested commutators with the observable, which can be reused in quantum simulation contexts beyond the semiclassical regime.
  • Potentials with Coulomb-type singularities are outside Assumption 1, and the paper notes that first-order Trotter for such potentials has at most a $1/4$ convergence rate, so the uniform-in-$h$ result does not extend to them.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same height-width counting should extend to time-dependent potentials $V(t,x)$, since the commutator estimates do not use time-independence; the paper lists this as future work, but the mechanism already carries it.
  • The local-to-global step is the part to watch: Theorem 3.1 controls one step starting from $O$, while the global claim needs the approximate Heisenberg flow to keep the observable in the polynomial class after $n$ steps. The text does not prove that invariance, so a decisive test is whether the global error stays flat in $h$ at fixed $\Delta t$ for $t$ well beyond one step.
  • If the $h$-independent bound holds, observable-oriented quantum algorithms for semiclassical dynamics could be quoted with cost $\operatorname{poly}(1/\epsilon, \operatorname{polylog}(1/h))$ rather than $\operatorname{poly}(h^{-1})$, shifting attention from wavefunction fidelity to expectation-value accuracy.
  • Non-polynomial observables with Taylor coefficients of factorial decay are also covered by the commutator estimates, a point the paper makes in passing; this widens the practical class of quantities that can be computed without shrinking the time step.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper claims a uniform-in-h error bound for observables under p-th order Trotter-Suzuki splitting of the semiclassical Schrödinger equation. For polynomial observables O = Σ y_m(x) h^m ∂_x^m and discrete (finite-difference or spectral) Hamiltonians A + B, Theorem 2.1 asserts ||T_{p,n}(Δt) − T(t)|| ≤ C Δt^p after n = t/Δt steps, with C independent of h^{−1}. The proof has three parts: (i) a Taylor-expansion analysis of the one-step exact and Trotter dynamics (Section 3) reducing the local error to (p+1)-fold nested commutators of A, B with O; (ii) height-width lemmas in a Lie-algebra setting (Section 4) showing that those commutators have norm O(1) uniformly in h for both spectral and finite-difference discretizations; (iii) the assertion that the global n-step bound follows from the one-step bound. Numerical experiments confirm the predicted Δt^p convergence and h-independence of the observable error, contrasting with the h^{−1} scaling of the unitary error. My assessment is that parts (i) and (ii) are sound, while part (iii) is an unproved step that is load-bearing for the main theorem.

Significance. If substantiated, the claimed result is significant: it would extend the uniform-in-h observable error bounds of [12] from first/second order to arbitrary even order, with an elementary proof that bypasses the microlocal machinery used there, and it would justify h-independent Trotter step sizes for observable accuracy in the semiclassical regime (Corollary 2.3). The paper's strengths are real: the one-step commutator expansion in Section 3 is careful and self-contained; the height-width calculus of Section 4 is elegant and appears correct; and the numerical experiments probe the claimed scalings directly (including the contrast between O(h^{-1}) unitary errors and h-independent observable errors) rather than fitting parameters. The observable class (2.6) is explicit, and limitations (Coulomb potentials, even orders, one space dimension) are acknowledged. However, the qualification 'if substantiated' is essential: the passage from the one-step bound to the n-step bound is not proved, and that is precisely the point at which the semiclassical accumulation behavior is determined.

major comments (3)
  1. [§2.2 (after Eq. (2.7)); §3.3; Theorems 3.1, 4.4, 4.10] The global claim (2.7) is not derived from the local estimate. Theorem 2.1 asserts that it 'follows from the local error bound' (text after (2.7)), but no telescoping argument is given. Writing E(X) = e^{−iHΔt} X e^{iHΔt} and F(X) = U X U^†, the standard identity E^n O − F^n O = Σ_{j=0}^{n−1} E^j (E − F) F^{n−1−j} O shows that the n-step error requires the one-step local error bound (3.14) for observables of the form U^{n−1−j} O (U^†)^{n−1−j} (or, in the other ordering, for e^{iHjΔt} O e^{−iHjΔt}), uniformly in j = 0, …, n−1. These evolved observables are not finite polynomials of the form (2.6): in the finite-difference setting they are generically full N×N matrices outside the algebra eL_h of §4.2, and in the spectral setting they are not finite-order polynomials in ∂_x. Theorems 4.4 and 4.10 bound only nested commutators of the original polynomial O with A and B; no invariance of the height-width class under the exact or approximate Heisenberg flow is stated or proved. The closing sentence of Section 3.3 ('The uniformity in h reduces to proving that the commutators are independent of h') addresses only the fixed-observable local commutators. Since n = t/Δt grows as Δt → 0, this gap is exactly where the claimed O(Δt^p) accumulation must be established, and the stress-test concern about this transition is confirmed on reading the manuscript. A repair requires either a class-invariance lemma for evolved observables with h-uniform constants or a different global argument (e.g., expanding each evolved observable in a Taylor series about time zero and controlling the terms via Theorem 4.10 with explicit, t-dependent constants).
  2. [§2.2, Proposition 2.5 (Eq. (2.8))] Proposition 2.5 is stated as a global bound, ||U_p^{†t} O U_p^t − e^{iHt} O e^{−iHt}|| ≤ C_p β_comm t^{p+1}, but no proof is given, and as printed the statement cannot serve as the bridge between Theorem 3.1 and Theorem 2.1. If t denotes a single step, (2.8) is merely the one-step estimate (3.14) restated with Δt replaced by t. If t denotes the total evolution time, the right-hand side is independent of the number of steps and cannot produce the O(Δt^p) = O(t^p n^{−p}) rate claimed in Theorem 2.1, which vanishes as n → ∞; at most it gives a loose bound of the form C_p β_comm t^{p+1} after summing n one-step estimates, and that summation still requires the evolved-observable control identified in my first comment. Moreover, the sentence introducing the proposition ('As a byproduct of the proof of our main theorem') is inaccurate because the main theorem's proof is exactly what is missing the associated step. The proposition should either be proved with correct n-dependence and a correct statement, or removed from the paper.
  3. [§4.2, Theorem 4.10; Theorem 2.1] The uniform-in-h statements are missing the required relation between the grid size and h. In the fully discrete setting, 'height' counts powers of N, and the proof of Theorem 4.10 bounds each term of a nested commutator W by C h^{wd} N^{ht} with ht ≤ wd, hence by C (hN)^{ht}; the conclusion ∥W∥ = O(1) as h → 0 therefore requires N = O(h^{−1}) (equivalently Δx ≳ h). For over-resolved grids with N ≫ h^{−1}, the bound degrades as (hN)^{ht}, and for simple commutators (e.g., [−(h/2)D_2, Y_1 h D] ≈ h^2 [D_2, Y_1] D, whose norm is of order h^2 N^2) the norms genuinely diverge as h → 0, so the theorem as stated is false without the grid-size hypothesis. For spectral discretizations the paper states the needed Δx = O(h) (Section 4.1), but Theorem 4.10 and Theorem 2.1 omit the analogous condition for finite differences; the claim 'uniformly in h ∈ (0,1]' should be accompanied by an explicit N–h scaling assumption.
minor comments (4)
  1. [§5, Figure 2] The numerical experiments include a '1st-order' Trotter scheme, but the theory in Sections 2–3 treats only even-order Suzuki methods (p = 2k in (2.2)); first-order uniform bounds are covered by prior work [12] but not by the present paper's theorems. Please either extend the theory to first order or restrict the experiments to even orders.
  2. [§3.2, Eqs. (3.9)–(3.10)] The remainder formula (3.9) and the definition of α_comm in (3.10) have an index inconsistency: (3.9) sums over sequences q_1 + … + q_k with k depending on the term, while the line preceding (3.10) writes Σ_{q_1+…+q_l = p+1}, and (3.10) then uses k without defining the length of the sequence of H_j's. Please harmonize the notation and specify the multinomial coefficient conventions.
  3. [§4.1, remark after Theorem 4.4] The discussion of non-polynomial observables asserts that 'the largest norm of a grade-n commutator is bounded by n^m, and there are at most 2^{n−1} possible choices of grade-n nested commutators' without proof. Since this is presented as a remark rather than a theorem, please flag these growth estimates as heuristic, or provide a short derivation.
  4. [§1, informal main-result statement] The informal bound in the introduction, ||U_app^† O U_app − e^{iHt} O e^{−iHt}|| ≤ C n^{−p}, is stated with a constant C that is implicitly t-dependent, while Theorem 2.1 states ||T_{p,n}(Δt) − T(t)|| ≤ C Δt^p. Since n = t/Δt, these agree only if the introduction's C absorbs t^p; please reconcile the two statements to avoid confusion about what is being claimed.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the algebraic estimates are proved directly and the result is not fitted; self-citations are present but not load-bearing.

full rationale

The paper's central claims (Theorems 2.1, 3.1, 4.4, and 4.10) are derived from explicit Taylor expansions and direct commutator calculations in Sections 3 and 4 and Appendix A, not from fitting or from the authors' prior results. Assumption 2 defines the admissible polynomial observables, and the height/width lemmas are proved in the text (Lemma 4.1 and Appendix A; Lemmas 4.5-4.7 for finite differences) rather than imported from the cited works. The constants C depend on V, O, p, and t but are not calibrated to experiments; Section 5 only illustrates the proved scalings. Prior work by the authors, e.g. [12], is cited for context and for related methods, but the current argument does not invoke those results as load-bearing premises. The main caveat is that the paper states Theorem 2.1 follows from the one-step local error (Theorem 3.1) without spelling out the telescoping argument that controls the accumulated error over n steps for time-evolved observables. This is a potential correctness/completeness gap, not a circular reduction: the one-step bound is proved for the original observable O, and no fitted parameter or renamed input is used to obtain the global bound. Therefore the circularity score is low.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard Taylor expansion, the order conditions of Trotter-Suzuki formulas, the grid scaling Δx = O(h), and the smoothness assumptions on V and the observable. The most fragile item in the ledger is the unproven local-to-global accumulation assumption, which is not stated in the paper.

assumptions (6)
  • standard math Taylor expansion with integral remainder is valid for the (bounded, discretized) operators A, B, and O.
    Used throughout Section 3 to expand e^{iHs} O e^{-iHs} and the Trotterized version. For finite matrices this is standard.
  • domain assumption The p-th order Trotter-Suzuki formula (2.2) satisfies the classical order conditions up to order p, so all terms of order s^j, j≤p, cancel in the local error.
    This is the defining property of the Suzuki/Yoshida construction, cited to [31,32]. It is the backbone of the local error bound in Section 3.
  • domain assumption The spatial grid is chosen with Δx = O(h), so N = O(h^{-1}), implying that the discretized k-th derivative has operator norm O(h^{-k}) for spectral methods, and analogous N-growth for finite differences.
    Used in Section 4 to convert the height≤width property into a uniform-in-h norm bound (Theorems 4.4 and 4.10).
  • domain assumption Assumption 1: V ∈ S(1), i.e., smooth and bounded with all derivatives.
    Assumption 1 in Section 2.1; needed for the height-width algebra to yield h-independent constants.
  • domain assumption Assumption 2: the observable is a finite polynomial Σ y_m(x) h^m ∂_x^m with y_m ∈ S(1).
    Assumption 2 in Section 2.2; the theorem is restricted to this class.
  • ad hoc to paper The local-to-global error accumulation (telescoping sum) is valid, and the local error bound extends to all observables appearing in the sum.
    This is the missing step. The paper does not prove it; it is flagged as the weakest assumption. Without it, Theorem 2.1 does not follow from Theorem 3.1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uniform semiclassical observable error bound of Trotter-Suzuki splitting: a simple algebraic proof." pith.science (2026). https://pith.science/paper/MC3XTFAQ

@misc{pith2026250702783,
  author       = {Pith},
  title        = {Pith review of: Uniform semiclassical observable error bound of Trotter-Suzuki splitting: a simple algebraic proof},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MC3XTFAQ}},
  note         = {Machine review of arXiv:2507.02783}
}
abstract

Efficient simulation of the semiclassical Schr\"odinger equation has garnered significant attention in the numerical analysis community. While controlling the error in the unitary evolution or the wavefunction typically requires the time step size to shrink as the semiclassical parameter $h$ decreases, it has been observed -- and proved for first- and second-order Trotterization schemes -- that the error in certain classes of observables admits a time step size independent of $h$. In this work, we explicitly characterize this class of observables and present a new, simple algebraic proof of uniform-in-$h$ error bounds for arbitrarily high-order Trotterization schemes. Our proof relies solely on the algebraic structure of the underlying operators in both the continuous and discrete settings. Unlike previous analyses, it avoids Egorov-type theorems and bypasses heavy semiclassical machinery. To our knowledge, this is the first proof of uniform-in-$h$ observable error bounds for Trotterization in the semiclassical regime that relies only on algebraic structure, without invoking the semiclassical limit.

Figures

Figures reproduced from arXiv: 2507.02783 by the authors.

Figure 1
Figure 1. This figure illustrates the distinct proof strategy employed in our work compared to [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Log-log plots showing the convergence of errors with respect to the step size ∆ [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗
Figure 3
Figure 3. Log-log plots showing unitary and observable errors as a function of the semiclassical [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Log-log plot of commutators illustrating the scaling behavior of nested commutators with [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. High-order Magnus Expansion for Hamiltonian Simulation

    quant-ph 2025-09 conditional novelty 7.0 of 10

    Arbitrary-order Magnus expansion is shown to have commutator-scaling error bounds and a polynomial-cost quantum circuit, yielding a time-dependent Hamiltonian simulation algorithm with O~(αbar^{1+1/p} T^{1+1/p}/ε^{1/p...

Reference graph

Works this paper leans on

60 extracted references · 40 canonical work pages · cited by 1 Pith paper

  1. [12]

    Zurich Lectures in Advanced Mathematics

    Christian Lubich.From quantum to classical molecular dynamics: reduced models and nu- merical analysis. Zurich Lectures in Advanced Mathematics. European Mathematical Society (EMS), Z¨ urich, 2008. ISBN 978-3-03719-067-8. doi: 10.4171/067. URL https://doi.org/10. 4171/067

    doi:10.4171/067 2008

  2. [1]

    Introduction.Simulation of quantum dynamics has been a foundational motivation for the development of quantum computers [1], and remains one of the most promising applications. In particular, quantum algorithms for Hamiltonian simulation aim to efficiently approximate the time evolution of quantum systems, and are widely regarded as a core primitive in qu...

    arXiv 2026

  3. [2]

    Problem Setup and Main Result.In this section, we set up the problem by review- ing the semiclassical Schr¨ odinger operator and Trotterization, and then present the main uniform estimates we obtained. 2.1. Higher-Order T rotter F ormula and Semiclassical Notations.We revisit thep-th order Trotter-Suzuki decomposition applied to simulate the time evolutio...

    work page

  4. [3]

    The central result of this section is the following theorem, which provides a uniform-in-hupper bound on the local error for the semiclassical Schr¨ odinger operators

    T rotter Error for Time-Evolved Observables.In this section, we establish the local error bound for thep-th order Trotter formula when applied to time-evolved observables. The central result of this section is the following theorem, which provides a uniform-in-hupper bound on the local error for the semiclassical Schr¨ odinger operators. Theorem3.1 (Local...

    work page

  5. [4]

    For completeness, we begin by introducing the underlying Lie algebra in a spatially continuous setting and reviewing its key properties

    Algebraic Structure and Lemmas.In this section, we establish the algebraic estimates required for bounding the Trotter observable error for the semiclassical Schr¨ odinger operators. For completeness, we begin by introducing the underlying Lie algebra in a spatially continuous setting and reviewing its key properties. We then utilize this algebraic struct...

    work page

  6. [5]

    The primary purpose of this section is to illustrate the behavior of higher-order Trotter-Suzuki formulas together with spatial discretization

    Numerical Comparison.While our theoretical results provide rigorous error bounds for higher-order Trotter formulas, we present numerical simulations to further validate these findings. The primary purpose of this section is to illustrate the behavior of higher-order Trotter-Suzuki formulas together with spatial discretization. We consider the semiclassica...

    work page

  7. [6]

    Conclusion and Discussion.In this work, we present the first purely algebraic proof of uniform-in-hobservable error bounds for arbitrarily high-order Trotter–Suzuki formulas applied to the semiclassical Schr¨ odinger equation. By directly comparing the exact and Trotterized evolution of a polynomial observableO, we avoid any reliance on Egorov-type theore...

    work page

  8. [7]

    Simulating physics with computers.International Journal of Theoretical Physics, 21(6):467–488, 1982

    Richard P Feynman. Simulating physics with computers.International Journal of Theoretical Physics, 21(6):467–488, 1982. doi: 10.1007/BF02650179

    doi:10.1007/bf02650179 1982

Show all 60 references
  1. [8]

    D. W. Berry, G. Ahokas, R. Cleve, and B. C. Sanders. Efficient quantum algorithms for simulating sparse Hamiltonians.Commun. Math. Phys., 270(2):359–371, 2007. doi: 10.1007/ s00220-006-0150-x

    2007

  2. [9]

    Andrew M. Childs. On the relationship between continuous- and discrete-time quantum walk. Comm. Math. Phys., 294(2):581–603, 2010. ISSN 0010-3616. doi: 10.1007/s00220-009-0930-1. URL https://doi.org/10.1007/s00220-009-0930-1

    2010 doi

  3. [10]

    Robin. Kothari. Efficient simulation of Hamiltonians.MS thesis. University of Waterloo., 2010. URL https://uwspace.uwaterloo.ca/bitstream/handle/10012/5356/RobinKothariThesis.pdf

    2010

  4. [11]

    Computing quantum dynamics in the semiclassical regime.Acta Numer., 29:229–401, 2020

    Caroline Lasser and Christian Lubich. Computing quantum dynamics in the semiclassical regime.Acta Numer., 29:229–401, 2020. ISSN 0962-4929. doi: 10.1017/s0962492920000033. URL https://doi-org.libproxy.berkeley.edu/10.1017/s0962492920000033

    2020 doi

  5. [13]

    Hagedorn and Alain Joye

    George A. Hagedorn and Alain Joye. Mathematical analysis of Born-Oppenheimer approxi- mations. InSpectral theory and mathematical physics: a Festschrift in honor of Barry Simon ’s 60th birthday, volume 76 ofProc. Sympos. Pure Math., pages 203–226. Amer. Math. Soc., Prov- idenc...

    2007 doi

  6. [14]

    Quantum algorithms for escaping from saddle points.Quantum, 5:529, Aug 2021

    Chenyi Zhang, Jiaqi Leng, and Tongyang Li. Quantum algorithms for escaping from saddle points.Quantum, 5:529, Aug 2021. ISSN 2521-327X. doi: 10.22331/q-2021-08-20-529. URL https://doi.org/10.22331/q-2021-08-20-529. 22DI F ANG AND CONRAD QU

    2021 doi

  7. [15]

    Su, and Tongyang Li

    Yizhou Liu, Weijie J. Su, and Tongyang Li. On quantum speedups for nonconvex optimization via quantum tunneling walks.Quantum, 7:1030, June 2023. ISSN 2521-327X. doi: 10.22331/ q-2023-06-02-1030. URL http://dx.doi.org/10.22331/q-2023-06-02-1030

    2023 doi

  8. [16]

    Quantum Hamiltonian descent, 2023

    Jiaqi Leng, Ethan Hickman, Joseph Li, and Xiaodi Wu. Quantum Hamiltonian descent, 2023. arXiv:2303.01471

    2023 arXiv

  9. [17]

    Mathematical and computational methods for semiclassical Schr¨ odinger equations.Acta Numer., 20:121–209, 2011

    Shi Jin, Peter Markowich, and Christof Sparber. Mathematical and computational methods for semiclassical Schr¨ odinger equations.Acta Numer., 20:121–209, 2011. ISSN 0962-4929. doi: 10.1017/S0962492911000031. URL https://doi.org/10.1017/S0962492911000031

    2011 doi

  10. [18]

    Uniform observable error bounds of Trotter formulae for the semiclassical Schr¨ odinger equation.Multiscale Modeling & Simulation, 23(1):255–277, 2025

    Yonah Borns-Weil and Di Fang. Uniform observable error bounds of Trotter formulae for the semiclassical Schr¨ odinger equation.Multiscale Modeling & Simulation, 23(1):255–277, 2025. doi: 10.1137/23M1583922. URL https://doi.org/10.1137/23M1583922

    2025 doi

  11. [19]

    Quantum simulation in the semi-classical regime.Quantum, 6:739, June 2022

    Shi Jin, Xiantao Li, and Nana Liu. Quantum simulation in the semi-classical regime.Quantum, 6:739, June 2022. ISSN 2521-327X. doi: 10.22331/q-2022-06-17-739. URL http://dx.doi.org/ 10.22331/q-2022-06-17-739

    2022 doi

  12. [20]

    W. Bao, S. Jin, and P. A. Markowich. On time-splitting spectral approximations for the Schr¨ odinger equation in the semiclassical regime.J. Comput. Phys., 175(2):487–524, 2002

    2002

  13. [21]

    Markowich

    Weizhu Bao, Shi Jin, and Peter A. Markowich. Numerical study of time-splitting spectral discretizations of nonlinear Schr¨ odinger equations in the semiclassical regimes.SIAM Journal on Scientific Computing, 25(1):27–64, 2003. doi: 10.1137/S1064827501393253. URL https: //doi.o...

    2003 doi

  14. [22]

    Markowich

    Weizhu Bao, Dieter Jaksch, and Peter A. Markowich. Numerical solution of the Gross–Pitaevskii equation for Bose–Einstein condensation.Journal of Computational Physics, 187(1):318–342, 2003. ISSN 0021-9991. doi: https://doi.org/10.1016/S0021-9991(03)00102-5. URL https://www.sci...

    2003 doi

  15. [23]

    On Fourier time-splitting methods for nonlinear Schr¨ odinger equations in the semiclassical limit.SIAM J

    R´ emi Carles. On Fourier time-splitting methods for nonlinear Schr¨ odinger equations in the semiclassical limit.SIAM J. Numer. Anal., 51(6):3232–3258, 2013. ISSN 0036-1429. doi: 10.1137/120892416. URL https://doi.org/10.1137/120892416

    2013 doi

  16. [24]

    On Fourier time-splitting methods for nonlinear Schr¨ odinger equations in the semi-classical limit II

    R´ emi Carles and Cl´ ement Gallo. On Fourier time-splitting methods for nonlinear Schr¨ odinger equations in the semi-classical limit II. Analytic regularity.Numer. Math., 136(1):315–342,

  17. [25]

    On the approximation of the von Neumann equation in the semi-classical limit

    Francis Filbet and Fran¸ cois Golse. On the approximation of the von Neumann equation in the semi-classical limit. part i : numerical algorithm, 2024. URL https://arxiv.org/abs/2405.13436

    2024 arXiv

  18. [26]

    A poisson integrator for Gaussian wavepacket dynamics

    Erwan Faou and Christian Lubich. A poisson integrator for Gaussian wavepacket dynamics. Computing and Visualization in Science, 2004

    2004

  19. [27]

    Gaussian beam methods for the Schrodinger equation in the semi- classical regime: Lagrangian and Eulerian formulations.Commun

    S Jin, H Wu, and X Yang. Gaussian beam methods for the Schrodinger equation in the semi- classical regime: Lagrangian and Eulerian formulations.Commun. Math. Sci., 6(1):995–1020, 2008

    2008

  20. [28]

    Computing semiclassical quantum dy- namics with Hagedorn wavepackets.SIAM Journal on Scientific Computing, 31(4):3027–3041,

    Erwan Faou, Vasile Gradinaru, and Christian Lubich. Computing semiclassical quantum dy- namics with Hagedorn wavepackets.SIAM Journal on Scientific Computing, 31(4):3027–3041,

  21. [29]

    On the convergence of time splitting methods for quantum dynamics in the semiclassical regime.Found

    Fran¸ cois Golse, Shi Jin, and Thierry Paul. On the convergence of time splitting methods for quantum dynamics in the semiclassical regime.Found. Comput. Math., 21(3):613–647, June 2021. ISSN 1615-3375. doi: 10.1007/s10208-020-09470-z. URL https://doi.org/10.1007/ s10208-020-09470-z

    2021 doi

  22. [30]

    Effective approx- imation for the semiclassical Schr¨ odinger equation.Foundations of Computational Mathe- matics, 14(4):689–720, 2014

    Philipp Bader, Arieh Iserles, Karolina Kropielnicka, and Pranav Singh. Effective approx- imation for the semiclassical Schr¨ odinger equation.Foundations of Computational Mathe- matics, 14(4):689–720, 2014. doi: 10.1007/s10208-013-9182-8. URL https://doi.org/10.1007/ s10208-013-9182-8

    2014 doi

  23. [31]

    An efficient time-splitting method for the Ehrenfest dynamics.Multiscale Modeling & Simulation, 16(2):900–921, 2018

    Di Fang, Shi Jin, and Christof Sparber. An efficient time-splitting method for the Ehrenfest dynamics.Multiscale Modeling & Simulation, 16(2):900–921, 2018. doi: 10.1137/17M1112789. URL https://doi.org/10.1137/17M1112789. UNIFORM SEMICLASSICAL OBSER V ABLE ERROR BOUND OF TROTT...

    2018 doi

  24. [32]

    Asymptotic analysis of diabatic surface hopping algorithm in the adiabatic and non-adiabatic limits, 2022

    Zhenning Cai, Di Fang, and Jianfeng Lu. Asymptotic analysis of diabatic surface hopping algorithm in the adiabatic and non-adiabatic limits, 2022. URL https://arxiv.org/abs/2205. 02312

    2022

  25. [33]

    CRC press, 2025

    Sergio Blanes and Fernando Casas.A concise introduction to geometric numerical integration. CRC press, 2025

    2025

  26. [34]

    On the approximation of the von Neumann equation in the semi-classical limit

    Fran¸ cois Golse and Francis Filbet. On the approximation of the von Neumann equation in the semi-classical limit. part II : numerical analysis, 2025. URL https://arxiv.org/abs/2504.18177

    2025

  27. [35]

    Mathematical theory and numerical methods for Bose-Einstein condensation.Kinetic & Related Models, 6(1):1–135, 2013

    Weizhu Bao and Yongyong Cai. Mathematical theory and numerical methods for Bose-Einstein condensation.Kinetic & Related Models, 6(1):1–135, 2013. ISSN 1937-5077. doi: 10.3934/ krm.2013.6.1. URL http://dx.doi.org/10.3934/krm.2013.6.1

    2013 doi

  28. [36]

    Quantitative observability for the Schr¨ odinger and Heisenberg equations: An optimal transport approach.Mathematical Models and Methods in Applied Sciences, 32(05):941–963, 2022

    Fran¸ cois Golse and Thierry Paul. Quantitative observability for the Schr¨ odinger and Heisenberg equations: An optimal transport approach.Mathematical Models and Methods in Applied Sciences, 32(05):941–963, 2022. doi: 10.1142/S021820252250021X. URL https://doi.org/10. 1142/S...

    2022 doi

  29. [37]

    Childs, Yuan Su, Minh C

    Andrew M. Childs, Yuan Su, Minh C. Tran, Nathan Wiebe, and Shuchen Zhu. Theory of Trotter error with commutator scaling.Phys. Rev. X, 11:011020, 2021. doi: 10.1103/PhysRevX. 11.011020

    2021 doi

  30. [38]

    Observable error bounds of the time-splitting scheme for quantum-classical molecular dynamics.SIAM Journal on Numerical Analysis, 61(1):26–44,

    Di Fang and Albert Tres Vilanova. Observable error bounds of the time-splitting scheme for quantum-classical molecular dynamics.SIAM Journal on Numerical Analysis, 61(1):26–44,

  31. [39]

    Efficient methods for linear Schr¨ odinger equation in the semiclassical regime with time-dependent potential

    Philipp Bader, Arieh Iserles, Karolina Kropielnicka, and Pranav Singh. Efficient methods for linear Schr¨ odinger equation in the semiclassical regime with time-dependent potential. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 472 (2193)...

    2016

  32. [40]

    M. Suzuki. General decomposition theory of ordered exponentials.Proc. Japan Acad., 69: 161–166, 1993. doi: 10.2183/pjab.69.161

    1993 doi

  33. [41]

    Construction of higher order symplectic integrators.Phys

    Haruo Yoshida. Construction of higher order symplectic integrators.Phys. Lett. A, 150(5-7): 262–268, 1990. ISSN 0375-9601,1873-2429. doi: 10.1016/0375-9601(90)90092-3. URL https: //doi.org/10.1016/0375-9601(90)90092-3

    1990 doi

  34. [42]

    Huyghebaert and H

    J. Huyghebaert and H. De Raedt. Product formula methods for time-dependent Schr¨ odinger problems.J. Phys. A, 23(24):5777–5793, 1990. doi: 10.1088/0305-4470/23/24/019

    1990 doi

  35. [43]

    Splitting and composition methods in the numerical integration of differential equations, 2008

    Sergio Blanes, Fernando Casas, and Ander Murua. Splitting and composition methods in the numerical integration of differential equations, 2008. URL https://arxiv.org/abs/0812.0377

    2008 arXiv

  36. [44]

    Zworski.Semiclassical Analysis

    M. Zworski.Semiclassical Analysis. Graduate studies in mathematics. ISBN 9780821883204. doi: 10.1090/gsm/138

    doi

  37. [45]

    St´ ephane Descombes and Mechthild Thalhammer. An exact local error representation of exponential operator splitting methods applied to linear Schr¨ odinger equations in the semi- classical regime.BIT Numerical Mathematics, 50(4):729–749, 2010. ISSN 1572-9125. doi: 10.1007/s10...

    2010 doi

  38. [46]

    Time-Dependent Hamiltonian Simulation via Magnus Expansion: Algorithm and Superconvergence.Communications in Mathematical Physics, 406 (6), May 2025

    Di Fang, Diyi Liu, and Rahul Sarkar. Time-Dependent Hamiltonian Simulation via Magnus Expansion: Algorithm and Superconvergence.Communications in Mathematical Physics, 406 (6), May 2025. ISSN 1432-0916. doi: 10.1007/s00220-025-05314-5. URL http://dx.doi.org/10. 1007/s00220-025-05314-5

    2025 doi

  39. [47]

    Algebraic theory for higher-order methods in computational quantum mechan- ics, 2015

    Pranav Singh. Algebraic theory for higher-order methods in computational quantum mechan- ics, 2015. URL https://arxiv.org/abs/1510.06896

    2015 arXiv

  40. [48]

    On the Trotter Error in Many-body Quantum Dynamics with Coulomb Potentials, 2025

    Di Fang, Xiaoxu Wu, and Avy Soffer. On the Trotter Error in Many-body Quantum Dynamics with Coulomb Potentials, 2025. URL https://arxiv.org/abs/2507.22707. arXiv:2507.22707

    2025

  41. [49]

    Magnus–Lanczos methods with sim- plified commutators for the Schr¨ odinger equation with a time-dependent potential.SIAM Journal on Numerical Analysis, 56(3):1547–1569, 2018

    Arieh Iserles, Karolina Kropielnicka, and Pranav Singh. Magnus–Lanczos methods with sim- plified commutators for the Schr¨ odinger equation with a time-dependent potential.SIAM Journal on Numerical Analysis, 56(3):1547–1569, 2018. doi: 10.1137/17M1149833. URL https://doi.org/1...

    2018 doi

  42. [50]

    Compact schemes for laser–matter in- teraction in Schr¨ odinger equation based on effective splittings of magnus expansion.Computer Physics Communications, 234:195–201, 2019

    Arieh Iserles, Karolina Kropielnicka, and Pranav Singh. Compact schemes for laser–matter in- teraction in Schr¨ odinger equation based on effective splittings of magnus expansion.Computer Physics Communications, 234:195–201, 2019. ISSN 0010-4655. doi: https://doi.org/10.1016/j...

    2019 doi

  43. [52]

    Time-dependent unbounded Hamiltonian simulation with vector norm scaling.Quantum, 5:459, may 2021

    Dong An, Di Fang, and Lin Lin. Time-dependent unbounded Hamiltonian simulation with vector norm scaling.Quantum, 5:459, may 2021. ISSN 2521-327X. doi: 10.22331/ q-2021-05-26-459. URL https://doi.org/10.22331/q-2021-05-26-459

    2021 doi

  44. [53]

    Ikeda, and Keisuke Fujii

    Kaoru Mizuta, Tatsuhiko N. Ikeda, and Keisuke Fujii. Explicit error bounds with commutator scaling for time-dependent product and multi-product formulas, 2024. https://arxiv.org/abs/2410.14243

    2024

  45. [54]

    Time-dependent Hamiltonian Simulation of Highly Os- cillatory Dynamics and Superconvergence for Schr¨ odinger Equation.Quantum, 6:690, April

    Dong An, Di Fang, and Lin Lin. Time-dependent Hamiltonian Simulation of Highly Os- cillatory Dynamics and Superconvergence for Schr¨ odinger Equation.Quantum, 6:690, April

  46. [57]

    Discrete superconvergence analysis for quantum magnus algorithms of unbounded hamiltonian simulation.Communications in Mathematical Physics, 407(2), January 2026

    Yonah Borns-Weil, Di Fang, and Jiaqi Zhang. Discrete superconvergence analysis for quantum magnus algorithms of unbounded hamiltonian simulation.Communications in Mathematical Physics, 407(2), January 2026. ISSN 1432-0916. doi: 10.1007/s00220-025-05531-y. URL http://dx.doi.org...

    2026 doi

  47. [59]

    Strong error bounds for Trotter and strang-splittings and their implications for quantum chem- istry.Physical Review Research, 6(4), November 2024

    Daniel Burgarth, Paolo Facchi, Alexander Hahn, Mattias Johnsson, and Kazuya Yuasa. Strong error bounds for Trotter and strang-splittings and their implications for quantum chem- istry.Physical Review Research, 6(4), November 2024. ISSN 2643-1564. doi: 10.1103/ physrevresearch....

    2024 doi

  48. [60]

    Convergence Rates for the Trotter Splitting for Unbounded Operators.Foundations of Computational Mathematics, September 2025

    Simon Becker, Niklas Galke, Lauritz van Luijk, and Robert Salzmann. Convergence Rates for the Trotter Splitting for Unbounded Operators.Foundations of Computational Mathematics, September 2025. ISSN 1615-3383. doi: 10.1007/s10208-025-09730-w. URL http://dx.doi.org/ 10.1007/s10...

    2025 doi

  49. [2009]

    URL https://doi.org/10.1137/080729724

    doi: 10.1137/080729724. URL https://doi.org/10.1137/080729724

    doi

  50. [2017]

    doi: 10.1007/s00211-016-0841-y

    ISSN 0029-599X. doi: 10.1007/s00211-016-0841-y. URL https://doi.org/10.1007/ s00211-016-0841-y

    doi

  51. [2022]

    doi: 10.22331/q-2022-04-15-690

    ISSN 2521-327X. doi: 10.22331/q-2022-04-15-690. URL https://doi.org/10.22331/ q-2022-04-15-690

    2022 doi

  52. [2023]

    URL https://doi.org/10.1137/21M1462349

    doi: 10.1137/21M1462349. URL https://doi.org/10.1137/21M1462349

    doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.