arxiv: 2604.24319 · v1 · submitted 2026-04-27 · 🧮 math.NA · cs.NA· math.PR

Recognition: unknown

Strong convergence and temporal-spatial regularity for tamed Euler approximations of L\'evy-driven SDEs

Yan Ding , Sizhou Wu , Ying Zhang

Authors on Pith no claims yet

Pith reviewed 2026-05-08 02:05 UTC · model grok-4.3

classification 🧮 math.NA cs.NAmath.PR

keywords tamed Euler schemestrong convergencetemporal-spatial regularityLévy-driven SDEssuperlinear growthstability estimatescontinuity estimatesnumerical approximation

0 comments

The pith

A novel tamed Euler scheme converges strongly to solutions of Lévy-driven SDEs with superlinear coefficients and supplies temporal-spatial regularity estimates.

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

The paper develops a new tamed Euler-type numerical scheme tailored to stochastic differential equations driven by Lévy processes whose drift and diffusion grow faster than linearly. It proves strong convergence of the scheme to the true solution and then derives estimates that quantify how the approximation changes smoothly when the starting value, starting time, or evaluation time is varied. These include a stability result that holds for any fixed time step and a continuity result that holds when the time step is allowed to shrink to zero. A reader would care because such schemes enable reliable pathwise simulation of processes that arise in finance, physics, and biology, where ordinary Euler methods lose control once coefficients become superlinear.

Core claim

We introduce a novel tamed Euler-type scheme for Lévy-driven SDEs with superlinearly growing drift and diffusion coefficients and establish its strong convergence. We then derive temporal-spatial regularity estimates with respect to the initial value, the initial time, and the evaluation time. In particular, we obtain a stability estimate for fixed step size and a corresponding continuity estimate in the vanishing step-size regime.

What carries the argument

The tamed Euler-type scheme, which augments the standard Euler step with a taming function that restores moment bounds and Lipschitz-type control while preserving the underlying Lévy increments.

If this is right

The scheme supplies a practical way to generate approximate sample paths with a quantifiable strong error that remains controlled even when coefficients grow superlinearly.
Temporal-spatial regularity yields uniform error bounds that are stable under perturbations of the initial data or time horizon.
The fixed-step stability estimate permits reliable long-time integration without step-size reduction at every step.
The vanishing-step continuity estimate justifies the use of the scheme inside adaptive or multilevel Monte Carlo algorithms.
Numerical experiments in the paper confirm that the observed convergence rates align with the theoretical predictions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same taming idea may carry over to SDEs driven by other jump processes or to equations with delay.
Regularity estimates of this form could be used to bound the bias introduced by time-discretization inside parameter-estimation procedures.
If the taming function can be chosen in a data-driven way, the method might extend to partially observed or high-dimensional systems.
The continuity result suggests that the scheme could serve as a building block for hybrid deterministic-stochastic solvers.

Load-bearing premise

The drift and diffusion admit a taming function that keeps moments bounded and restores enough regularity for the convergence and stability proofs to go through.

What would settle it

An explicit SDE satisfying the stated growth conditions for which the proposed tamed scheme produces unbounded moments or fails to converge strongly in L^p for any p greater than one.

Figures

Figures reproduced from arXiv: 2604.24319 by Sizhou Wu, Yan Ding, Ying Zhang.

**Figure 1.** Figure 1: Simulation results for the tamed Euler scheme (18) corresponding to Theorem 1 (i)–(iii). For brevity, we denote Y x,δ,ϵ,M tn and Y x,δ,ϵ, ˜ M tn by Y x,δ tn and Y x,δ ˜ tn , respectively. for different initial gaps |x − x˜| ∈ {10−10 , 10− 68 7 , 10− 66 7 , ..., 10− 60 7 , 10− 58 7 , 10−8} with N = 1000. The resulting log-log plot shows an approximately linear dependence of the L 2 distance on the initial c… view at source ↗

**Figure 2.** Figure 2: Simulation results for the tamed Euler scheme (21) corresponding to Theorem 1 (i)–(iii). 4.2. 1D SDE in financial model. Consider the one-dimensional L´evy-driven SDE with 3/2-type volatility (20) dXt = aXt−(b − |Xt−|) dt + c|Xt−| 3 2 dWt + Xt− Z R z π˜(dz, dt), for all t ∈ [0, 1], with parameters a, b, c ≥ 0 and initial value X0 = 2. We use the L´evy measure defined in (♢). Proposition 2. The coefficients… view at source ↗

**Figure 3.** Figure 3: A illustration for the case distinction. The points {ti , ti+1, ti+2} indicate grid points (drawn by ×). Case (iii) we treat the complementary situation where ˜s is a grid time, i.e., ˜s ∈ δ([0, T]), and again apply the disintegration theorem and the Markov property on grid points. Then, Lemmas 11 and 12 give (42). Finally, Case (iv) reduces the general situation of s ≤ s˜ to the previously established cas… view at source ↗

read the original abstract

We study the temporal-spatial regularity properties of tamed Euler approximations for L\'evy-driven SDEs with superlinearly growing drift and diffusion coefficients. We first introduce a novel tamed Euler-type scheme and establish its strong convergence. We then derive temporal-spatial regularity estimates with respect to the initial value, the initial time, and the evaluation time. In particular, we obtain a stability estimate for fixed step size and a corresponding continuity estimate in the vanishing step-size regime. Numerical experiments are presented to support the theoretical results.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper gives a tamed Euler scheme for Lévy-driven SDEs with superlinear coefficients, proves strong convergence, and adds temporal-spatial regularity estimates.

read the letter

The main thing to know about this paper is that it presents a novel tamed Euler-type scheme for Lévy-driven SDEs where the drift and diffusion have superlinear growth, establishes strong convergence of the approximations, and then derives temporal and spatial regularity estimates. The new part is applying taming to control the growth in the presence of Lévy jumps and getting those regularity results for the scheme. This seems useful because standard Euler methods can fail or diverge with superlinear terms, and the regularity gives ways to bound errors with respect to initial data and times. The paper does well in outlining the scheme and using the taming to get uniform moment bounds, which then lead to the convergence and the stability/continuity estimates. Including numerical experiments is a plus for supporting the theory. One soft spot is that everything hinges on the existence of a suitable taming function that preserves the needed properties, and while the full paper states the assumptions clearly per the stress test, those conditions might not be easy to check or satisfy in every application. The proofs are not visible in the abstract, so the details of how the jumps are handled could have some technical subtleties. This is aimed at researchers in numerical analysis for stochastic processes with jumps. Someone working on simulation of such SDEs would find the convergence guarantees and regularity tools relevant. The paper shows clear thinking on the problem and engages the literature, so it deserves a serious referee. I would recommend sending it for peer review.

Referee Report

0 major / 3 minor

Summary. The manuscript introduces a novel tamed Euler-type scheme for Lévy-driven SDEs whose drift and diffusion coefficients exhibit superlinear growth. It first establishes strong convergence of the scheme to the true solution under suitable growth, Lipschitz-type, and Lévy-measure integrability assumptions. It then derives temporal-spatial regularity estimates for the approximations with respect to the initial value, initial time, and evaluation time, including a stability bound for fixed step size h and a continuity estimate as h → 0. Numerical experiments are included to illustrate the theoretical findings.

Significance. If the stated assumptions on the coefficients and Lévy measure hold and the proofs close, the work supplies useful convergence and regularity tools for numerical simulation of jump-driven SDEs with superlinear coefficients. The combination of an explicit taming construction, strong convergence, and temporal-spatial regularity estimates (stability plus vanishing-step continuity) is a concrete advance over existing tamed schemes that typically stop at convergence without the full regularity picture.

minor comments (3)

The precise definition of the taming function (its growth control and preservation of moment bounds) is central; placing its explicit form and the full list of standing assumptions in a single early subsection would improve readability.
In the numerical section, the specific Lévy measures, truncation levels, and parameter values used in the experiments should be stated explicitly so that the reported error plots can be reproduced.
Notation for the compensated Poisson random measure and the compensated small-jump integral should be introduced once and used consistently; occasional re-definition of symbols interrupts the flow.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary of our manuscript and for recommending minor revision. The referee's description accurately reflects the main results on the novel tamed Euler scheme, strong convergence, and temporal-spatial regularity estimates for Lévy-driven SDEs with superlinear coefficients. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The derivation introduces a tamed Euler scheme for Lévy-driven SDEs under explicit growth/Lipschitz conditions on coefficients and the taming function, then proves strong convergence via standard moment estimates and Gronwall-type arguments before obtaining temporal-spatial regularity. These steps use classical stochastic-analysis tools (e.g., Itô formula, Burkholder-Davis-Gundy inequalities) that are independent of the target results; no quantity is defined in terms of itself, no fitted parameter is relabeled as a prediction, and no load-bearing premise reduces to a self-citation or ansatz smuggled from prior work by the same authors. The manuscript is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The central claims rest on standard existence/uniqueness results for Lévy SDEs and on the existence of a suitable taming function; no free parameters or new entities with independent evidence are introduced in the abstract.

axioms (2)

domain assumption The Lévy-driven SDE admits a unique strong solution under the stated superlinear growth conditions
Required for the approximation to be compared against a well-defined limit process.
domain assumption The taming function preserves the necessary moment bounds and allows the usual Itô calculus estimates to go through
Central to both the convergence proof and the regularity estimates.

invented entities (1)

Tamed Euler-type scheme no independent evidence
purpose: Numerical approximation that remains stable for superlinear coefficients
The scheme is introduced in the paper; no independent evidence outside the work is provided.

pith-pipeline@v0.9.0 · 5384 in / 1243 out tokens · 61366 ms · 2026-05-08T02:05:29.984350+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 1 canonical work pages

[1]

Applebaum.L´ evy processes and stochastic calculus, volume 116 ofCambridge Studies in Advanced Mathemat- ics

D. Applebaum.L´ evy processes and stochastic calculus, volume 116 ofCambridge Studies in Advanced Mathemat- ics. Cambridge University Press, Cambridge, second edition, 2009

2009
[2]

Baldeaux and A

J. Baldeaux and A. Badran. Consistent modelling of VIX and equity derivatives using a 3/2 plus jumps model. Appl. Math. Finance, 21(4):299–312, 2014

2014
[3]

O. E. Barndorff-Nielsen, T. Mikosch, and S. I. Resnick.L´ evy Processes: Theory and Applications. Springer Science & Business Media, 2012

2012
[4]

Cont and P

R. Cont and P. Tankov.Financial modelling with jump processes. Chapman & Hall/CRC Financial Mathematics Series. Chapman & Hall/CRC, Boca Raton, FL, 2004

2004
[5]

Dellacherie and P.-A

C. Dellacherie and P.-A. Meyer.Probabilities and potential. B, volume 72 ofNorth-Holland Mathematics Studies. North-Holland Publishing Co., Amsterdam, 1982. Theory of martingales, Translated from the French by J. P. Wilson

1982
[6]

W. E, M. Hutzenthaler, A. Jentzen, and T. Kruse. Multilevel Picard iterations for solving smooth semilinear parabolic heat equations.Partial Differ. Equ. Appl., 2(6):Paper No. 80, 31, 2021

2021
[7]

Goard and M

J. Goard and M. Mazur. Stochastic volatility models and the pricing of VIX options.Math. Finance, 23(3):439– 458, 2013

2013
[8]

Gonon and C

L. Gonon and C. Schwab. Deep ReLU neural networks overcome the curse of dimensionality for partial integrod- ifferential equations.Anal. Appl. (Singap.), 21(1):1–47, 2023

2023
[9]

Gy¨ ongy and N

I. Gy¨ ongy and N. V. Krylov. On stochastic equations with respect to semimartingales. I.Stochastics, 4(1):1–21, 1980/81

1980
[10]

Gy¨ ongy and S

I. Gy¨ ongy and S. Wu. On Itˆ o formulas for jump processes.Queueing Syst., 98(3-4):247–273, 2021

2021
[11]

Hutzenthaler and A

M. Hutzenthaler and A. Jentzen. Numerical approximations of stochastic differential equations with non-globally Lipschitz continuous coefficients.Mem. Amer. Math. Soc., 236(1112):v+99, 2015

2015
[12]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, and P. E. Kloeden. Strong and weak divergence in finite time of Euler’s method for stochastic differential equations with non-globally Lipschitz continuous coefficients.Proc. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci., 467(2130):1563–1576, 2011

2011
[13]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, and P. E. Kloeden. Strong convergence of an explicit numerical method for SDEs with nonglobally Lipschitz continuous coefficients.Ann. Appl. Probab., 22(4):1611–1641, 2012

2012
[14]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, and T. Kruse. Overcoming the curse of dimensionality in the numerical approxi- mation of parabolic partial differential equations with gradient-dependent nonlinearities.Found. Comput. Math., 22(4):905–966, 2022

2022
[15]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, T. Kruse, and T. A. Nguyen. Multilevel Picard approximations for high-dimensional semilinear second-order PDEs with Lipschitz nonlinearities.arXiv preprint arXiv:2009.02484, 2020

work page arXiv 2009
[16]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, T. Kruse, and T. A. Nguyen. Overcoming the curse of dimensionality in the numerical approximation of backward stochastic differential equations.J. Numer. Math., 31(1):1–28, 2023. 3Recall that 0/0:=0 by convention 51

2023
[17]

Hutzenthaler, A

M. Hutzenthaler, A. Jentzen, T. Kruse, T. A. Nguyen, and P. von Wurstemberger. Overcoming the curse of dimensionality in the numerical approximation of semilinear parabolic partial differential equations.Proc. Roy. Soc. A, 476(2244):20190630, 25, 2020

2020
[18]

Hutzenthaler and T

M. Hutzenthaler and T. Kruse. Multilevel Picard approximations of high-dimensional semilinear parabolic dif- ferential equations with gradient-dependent nonlinearities.SIAM J. Numer. Anal., 58(2):929–961, 2020

2020
[19]

Hutzenthaler and T

M. Hutzenthaler and T. A. Nguyen. Strong convergence rate of Euler-Maruyama approximations in temporal- spatial H¨ older-norms.J. Comput. Appl. Math., 413:Paper No. 114391, 17, 2022

2022
[20]

P. E. Kloeden and E. Platen.Numerical solution of stochastic differential equations, volume 23 ofApplications of Mathematics (New York). Springer-Verlag, Berlin, 1992

1992
[21]

Kumar and S

C. Kumar and S. Sabanis. On explicit approximations for L´ evy driven SDEs with super-linear diffusion coeffi- cients.Electron. J. Probab., 22:Paper No. 73, 19, 2017

2017
[22]

Kumar and S

C. Kumar and S. Sabanis. On Milstein approximations with varying coefficients: the case of super-linear diffusion coefficients.BIT, 59(4):929–968, 2019

2019
[23]

Mao.Stochastic differential equations and applications

X. Mao.Stochastic differential equations and applications. Horwood Publishing Limited, Chichester, second edition, 2008

2008
[24]

Mikulevicius and H

R. Mikulevicius and H. Pragarauskas. OnL p-estimates of some singular integrals related to jump processes. SIAM J. Math. Anal., 44(4):2305–2328, 2012

2012
[25]

Neufeld and S

A. Neufeld and S. Wu. Multilevel Picard approximation algorithm for semilinear partial integro-differential equations and its complexity analysis.Stoch. Partial Differ. Equ. Anal. Comput., 13(3):1220–1278, 2025

2025
[26]

S. Sabanis. A note on tamed Euler approximations.Electron. Commun. Probab., 18:1–10, 2013

2013
[27]

S. Sabanis. Euler approximations with varying coefficients: the case of superlinearly growing diffusion coefficients. Ann. Appl. Probab., 26(4):2083–2105, 2016

2083
[28]

Situ.Theory of stochastic differential equations with jumps and applications

R. Situ.Theory of stochastic differential equations with jumps and applications. Mathematical and Analytical Techniques with Applications to Engineering. Springer, New York, 2005

2005
[29]

C. H. Yuen, W. Zheng, and Y. K. Kwok. Pricing exotic discrete variance swaps under the 3/2-stochastic volatility models.Appl. Math. Finance, 22(5):421–449, 2015

2015
[30]

V. A. Zorich.Mathematical analysis. I. Universitext. Springer-Verlag, Berlin, Russian edition, 2004. Financial Technology Thrust, Society Hub, The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China Email address:yding632@connect.hkust-gz.edu.cn School of Mathematics, Shanghai University of Finance and Economics, Shanghai, China E...

2004