The backward problem for a multi-term time-fractional diffusion equation

Damir Shamuratov; Ravshan Ashurov

arxiv: 2604.09239 · v1 · submitted 2026-04-10 · 🧮 math.AP

The backward problem for a multi-term time-fractional diffusion equation

Ravshan Ashurov , Damir Shamuratov This is my paper

Pith reviewed 2026-05-10 17:54 UTC · model grok-4.3

classification 🧮 math.AP

keywords backward problemmulti-term fractional diffusionMittag-Leffler functionill-posed problemsmoothing propertyexistence and uniquenessconditional stabilityinverse problem

0 comments

The pith

With sufficiently smooth terminal data, the backward problem for a multi-term time-fractional diffusion equation has a unique stable solution that belongs to the domain of the spatial operator for every positive time.

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

The paper establishes that the backward problem for a multi-term time-fractional diffusion equation, which is typically ill-posed, becomes well-posed when the data at the final time is sufficiently smooth. Under this condition the initial data can be uniquely recovered in a stable manner. The argument relies on a precise asymptotic description of the multinomial Mittag-Leffler function that appears in the solution representation. The work further shows that any such solution instantly lies in the domain of the elliptic operator A, which is the strongest smoothing property possible for the equation.

Core claim

For the backward problem of recovering the initial condition from data at time T in a multi-term time-fractional diffusion equation, sufficiently smooth terminal data guarantee existence, uniqueness, and stability of the solution. The same data class also forces the solution to belong to the domain of the spatial operator A for every positive time. Conditional stability holds when an a priori bound on the initial datum is imposed. The proofs rest on the asymptotic expansion of the multinomial Mittag-Leffler function that governs the solution formula.

What carries the argument

The asymptotic characterization of the multinomial Mittag-Leffler function, which controls the denominator in the solution representation and enables the stability and smoothing estimates.

If this is right

Existence and uniqueness hold for the inverse problem under a smoothness assumption on the final data.
The recovered solution satisfies the strongest possible smoothing property by entering the domain of A immediately after t=0.
Conditional stability is restored once an a priori bound is placed on the unknown initial datum.
The same argument applies to any finite number of fractional time derivatives whose orders lie in (0,1).

Where Pith is reading between the lines

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

The same smoothing conclusion may hold for other linear fractional evolution equations whose solution kernels admit analogous asymptotic expansions.
Numerical schemes for the inverse problem could exploit the instant membership in D(A) to obtain higher-order approximations after a short time.
In physical models of anomalous diffusion the result suggests that initial states can be reconstructed reliably from later measurements provided those measurements are sufficiently regular.

Load-bearing premise

The multinomial Mittag-Leffler function must obey the precise asymptotic decay established in the authors' earlier work.

What would settle it

A concrete counter-example in which the multinomial Mittag-Leffler function fails to satisfy the claimed asymptotic bounds for large positive arguments would falsify the stability and smoothing conclusions.

read the original abstract

This paper is devoted to the investigation of the backward problem for a multi-term time-fractional diffusion equation. Backward problems for fractional diffusion equations are typically studied using regularization methods due to their ill-posedness in the sense of Hadamard; that is, a small change in u(T) may lead to large changes in the initial data. Nevertheless, we show that if sufficiently smooth current data are considered, then the solution exists, is unique, and is stable. A principal difficulty in the analysis of the backward problem stems from the structure of the solution, in which the multinomial Mittag-Leffler function appears in the denominator. Accordingly, a precise characterization of the asymptotic behavior of this function is required. Such asymptotic properties are nontrivial and have been rigorously established in the authors' recent work, which serves as a fundamental basis for the present study. In addition, we investigate the conditional stability of the backward problem. It is shown that, although the problem is ill-posed in general, stability can be restored under an appropriate a priori bound imposed on the initial data. The main novelty of the paper lies in proving the best smoothing property of the solution, showing that it belongs to the domain of the operator A for any positive time.

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

The paper proves well-posedness and smoothing for the multi-term backward fractional diffusion problem when data is smooth, but depends heavily on the authors' prior asymptotics for the multinomial Mittag-Leffler function.

read the letter

The punchline is that this paper proves existence, uniqueness, stability, and a strong smoothing property for the backward problem in the multi-term time-fractional diffusion equation, but only when the terminal data is sufficiently smooth, and it does so by relying on the authors' prior asymptotic results for the multinomial Mittag-Leffler function. The new contribution is the extension to the multi-term case and especially the demonstration that the solution belongs to the domain of the spatial operator A for every positive time. They handle the ill-posedness by showing conditional stability under an a priori bound on the initial data. The explicit solution representation makes the analysis direct once the asymptotics are in hand, and the paper does a good job organizing the estimates around those asymptotics. The soft spot is the dependence on the earlier work for the asymptotic expansion of the Mittag-Leffler function. Since the key decay rates and operator norm bounds come straight from that paper, any restriction in the admissible angles or any missing logarithmic terms would affect the conclusions here. The current manuscript does not re-establish those asymptotics, so it is not fully self-contained on that point. The precise meaning of sufficiently smooth data also needs to be checked against the function spaces used in the proofs. This is for people working on inverse problems for fractional equations, particularly those interested in multi-term models. A reader who needs the smoothing estimate or the conditional stability result would find it directly applicable. I think it deserves a serious referee. The thinking is clear, the results are stated precisely, and the contribution is a useful incremental step in the literature. Recommendation: Send it to peer review.

Referee Report

2 major / 3 minor

Summary. The manuscript investigates the backward problem for a multi-term time-fractional diffusion equation. It claims that, for sufficiently smooth terminal data at t=T, the solution exists, is unique, and is stable. It further establishes conditional stability under an a priori bound on the initial data and proves the smoothing property that the recovered solution belongs to the domain of the spatial operator A for every t>0. The analysis proceeds from an explicit solution representation whose denominator involves the multinomial Mittag-Leffler function, whose asymptotic properties (as |z|→∞ in suitable sectors) are imported from the authors' prior work.

Significance. If the central claims hold, the paper contributes a rigorous well-posedness theory for the backward multi-term fractional diffusion problem, showing that smoothness of the current data restores existence, uniqueness, and stability while also delivering the optimal smoothing effect of membership in D(A) for t>0. The explicit representation and the precise use of multinomial Mittag-Leffler asymptotics constitute a technical strength that extends single-term results and supplies falsifiable decay rates for future numerical validation.

major comments (2)

[§3] §3 (solution representation): the explicit formula for the initial datum (presumably Eq. (3.2) or equivalent) places the multinomial Mittag-Leffler function in the denominator; all subsequent operator-norm bounds in §4 and §5 are derived from its asymptotic expansion as |z|→∞. Because this expansion is cited verbatim from the authors' earlier paper without restating the admissible sector, the range of the fractional orders α_k, or any logarithmic corrections, the load-bearing estimates for existence, uniqueness, and the D(A) membership claim cannot be verified from the present manuscript alone.
[§5] §5 (conditional stability and smoothing): the proof that u(t) ∈ D(A) for all t>0 and the conditional stability estimate both rely on the decay rate furnished by the imported asymptotics. The manuscript does not quantify how the constants in these estimates depend on the smoothness class of the terminal data or on the specific multi-term coefficients; any restriction in the prior asymptotic result (e.g., sector angle or parameter range) would directly invalidate the claimed stability and smoothing statements.

minor comments (3)

The abstract and introduction use the phrase 'best smoothing property' without a brief comparison to the single-term case or a reference to the precise Sobolev or domain norm in which the improvement is measured.
[Preliminaries] Notation: the multi-term orders α_k and coefficients a_k are introduced without an explicit list of standing assumptions (e.g., 0<α_k<1, a_k>0) in the preliminaries section; this should be added for clarity.
[References] The reference list should include the exact citation details and page numbers for the authors' prior work on the multinomial Mittag-Leffler asymptotics so that readers can immediately locate the sector conditions.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive suggestions. The two major comments both concern the self-contained presentation of the asymptotic properties of the multinomial Mittag-Leffler function that underpin the well-posedness, stability, and smoothing results. We address each point below and will revise the manuscript accordingly.

read point-by-point responses

Referee: [§3] §3 (solution representation): the explicit formula for the initial datum (presumably Eq. (3.2) or equivalent) places the multinomial Mittag-Leffler function in the denominator; all subsequent operator-norm bounds in §4 and §5 are derived from its asymptotic expansion as |z|→∞. Because this expansion is cited verbatim from the authors' earlier paper without restating the admissible sector, the range of the fractional orders α_k, or any logarithmic corrections, the load-bearing estimates for existence, uniqueness, and the D(A) membership claim cannot be verified from the present manuscript alone.

Authors: We agree that the current citation leaves the key asymptotic hypotheses implicit. In the revised manuscript we will insert a short, self-contained statement (approximately one paragraph) that records the precise sector in which the expansion holds, the admissible range of the multi-term orders α_k, and the form of any logarithmic corrections. This addition will allow the subsequent operator-norm bounds to be checked directly from the present paper. revision: yes
Referee: [§5] §5 (conditional stability and smoothing): the proof that u(t) ∈ D(A) for all t>0 and the conditional stability estimate both rely on the decay rate furnished by the imported asymptotics. The manuscript does not quantify how the constants in these estimates depend on the smoothness class of the terminal data or on the specific multi-term coefficients; any restriction in the prior asymptotic result (e.g., sector angle or parameter range) would directly invalidate the claimed stability and smoothing statements.

Authors: We will augment the statements of the conditional stability and smoothing theorems to display the explicit dependence of the constants on the smoothness norm of the terminal datum and on the vector of fractional orders. We will also add a brief remark confirming that all estimates remain valid inside the parameter regime already assumed for the problem (i.e., the sector and order restrictions stated in the new paragraph of §3). These clarifications will remove any ambiguity about the scope of the results. revision: yes

Circularity Check

1 steps flagged

Key asymptotic estimates for multinomial Mittag-Leffler function imported from authors' prior work

specific steps

self citation load bearing [Abstract]
"a precise characterization of the asymptotic behavior of this function is required. Such asymptotic properties are nontrivial and have been rigorously established in the authors' recent work, which serves as a fundamental basis for the present study."

The solution formula places the multinomial Mittag-Leffler function in the denominator; all operator-norm bounds, decay rates, and well-posedness arguments (including the smoothing property that the recovered solution lies in D(A) for t>0) are obtained from its asymptotic expansion. These expansions are imported verbatim from the authors' earlier paper rather than proved here, so the load-bearing estimates reduce directly to the self-cited result.

full rationale

The paper establishes existence, uniqueness, stability, conditional stability, and the smoothing property (solution in D(A) for all t>0) for the backward problem via an explicit solution representation whose estimates require precise asymptotics of the multinomial Mittag-Leffler function E_{α,β}(z) as |z|→∞ in suitable sectors. These asymptotics are not re-derived but explicitly cited as the 'fundamental basis' from the authors' recent prior work. This creates a self-citation load-bearing dependence for the central claims, though the application to the backward problem and novelty in the smoothing proof retain independent content. No self-definitional, fitted-prediction, or renaming patterns appear; the derivation is otherwise self-contained once the cited asymptotics are granted.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims depend on the validity of the asymptotic characterization from the prior paper by the same authors.

axioms (1)

domain assumption Asymptotic properties of the multinomial Mittag-Leffler function as established in the authors' recent work
This is invoked as the basis for analyzing the solution structure in the backward problem.

pith-pipeline@v0.9.0 · 5519 in / 1103 out tokens · 29625 ms · 2026-05-10T17:54:35.825944+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages

[1]

Res., vol

Hatano, Y., Hatano, N.: Dispersive transport of ions in column experiments: An expla- nation of long-tailed profiles.Water Resour. Res., vol. 34, no. 5, pp. 1027–1033, 1998. https://doi.org/10.1029/98WR00214

work page doi:10.1029/98wr00214 1998
[2]

Res., vol

Schumer, R., Benson, D.A., Meerschaert, M.M., Baeumer, B.: Fractal mobile/immobile solute transport.Water Resour. Res., vol. 39, no. 10, Article ID 1296, 2003. https://doi.org/10.1029/2003WR002141

work page doi:10.1029/2003wr002141 2003
[3]

Elsevier, Amsterdam, 2006

Kilbas, A.A., Srivastava, H.M., Trujillo, J.J.:Theory and Applications of Fractional Differential Equations. Elsevier, Amsterdam, 2006

work page 2006
[4]

Springer, Berlin, 2006

Isakov, V.:Inverse Problems for Partial Differential Equations. Springer, Berlin, 2006

work page 2006
[5]

Theory and Applications

Kabanikhin, S.I.:Inverse and Ill-Posed Problems. Theory and Applications. De Gruyter, 2011

work page 2011
[6]

Sakamoto, K., Yamamoto, M.: Initial value/boundary value problems for fractional diffusion wave equations and applications to some inverse problems.J. Math. Anal. Appl., vol. 382, pp. 426–447, 2011. https://doi.org/10.1016/j.jmaa.2011.04.058

work page doi:10.1016/j.jmaa.2011.04.058 2011
[7]

Lincei Mat

Florida, G., Li, Z., Yamamoto, M.: Well-posedness for the backward problems in time for general time-fractional diffusion equation.Rend. Lincei Mat. Appl., vol. 31, pp. 593– 610, 2020. https://doi.org/10.4171/RLM/906

work page doi:10.4171/rlm/906 2020
[8]

Alimov, Sh., Ashurov, R.: On the backward problems in time for time-fractional subdiffusion equations.Fract. Differ. Calc., vol. 11, no. 2, pp. 203–217, 2021. https://doi.org/10.7153/fdc-2021-11-14 THE BACKW ARD PROBLEM FOR A MULTI-TERM . . . 13

work page doi:10.7153/fdc-2021-11-14 2021
[9]

Wei, T., Xian, J.: Variational method for a backward problem for a time- fractional diffusion equation.ESAIM: M2AN, vol. 53, no. 4, pp. 1223–1244, 2019. https://doi.org/10.1051/m2an/2019019

work page doi:10.1051/m2an/2019019 2019
[10]

Tuan, N.H., Le, H.N., Ngoc, T.B., Zhou, Y.: On a backward problem for non- linear fractional diffusion equations.Appl. Math. Lett., vol. 92, pp. 76–84, 2019. https://doi.org/10.1016/j.aml.2018.11.015

work page doi:10.1016/j.aml.2018.11.015 2019
[11]

Liu, J.J., Yamamoto, M.: A backward problem for the time-fractional diffusion equation. Appl. Anal., vol. 89, pp. 1769–1788, 2010. https://doi.org/10.1080/00036810903479731

work page doi:10.1080/00036810903479731 2010
[12]

Nonlinear Sci

Wen, J., Wang, Y.P., Wang, Y.X., Wang, Y.Q.: The quasi-reversibility regular- ization method for backward problem of the multi-term time-space fractional dif- fusion equation.Commun. Nonlinear Sci. Numer. Simul., Article ID 107848, 2024. https://doi.org/10.1016/j.cnsns.2024.107848

work page doi:10.1016/j.cnsns.2024.107848 2024
[13]

Math., vol

Sun, L., Wang, Y., Chang, M.: A fractional-order quasi-reversibility method to a back- ward problem for the multi-term time-fractional diffusion equation.Taiwanese J. Math., vol. 27, no. 6, pp. 1185–1210, 2023. https://doi.org/10.11650/tjm/230801

work page doi:10.11650/tjm/230801 2023
[14]

Li, Z., Liu, Y., Yamamoto, M.: Initial-boundary value problems for multi-term time- fractional diffusion equations with positive constant coefficients.Appl. Math. Comput., vol. 257, pp. 381–397, 2015. https://doi.org/10.1016/j.amc.2014.11.073

work page doi:10.1016/j.amc.2014.11.073 2015
[15]

https://doi.org/10.48550/arXiv.2603.01833

Ashurov, R., Shamuratov, D.: Inverse problem for a multi-term time- fractional diffusion equation with the Caputo derivatives.arXiv:2603.01833, 2026. https://doi.org/10.48550/arXiv.2603.01833

work page doi:10.48550/arxiv.2603.01833 2026
[16]

Springer, Berlin, 2014

Gorenflo, R., Kilbas, A.A., Mainardi, F., Rogosin, S.:Mittag-Leffler Functions, Related Topics and Applications. Springer, Berlin, 2014. Ravshan Ashurov: V. I. Romanovskiy Institute of Mathematics Uzbekistan Acad- emy of Sciences, University street 9, Tashkent–100174, Uzbekistan Email address:ashurovr@gmail.com Damir Shamuratov: V. I. Romanovskiy Institut...

work page 2014

[1] [1]

Res., vol

Hatano, Y., Hatano, N.: Dispersive transport of ions in column experiments: An expla- nation of long-tailed profiles.Water Resour. Res., vol. 34, no. 5, pp. 1027–1033, 1998. https://doi.org/10.1029/98WR00214

work page doi:10.1029/98wr00214 1998

[2] [2]

Res., vol

Schumer, R., Benson, D.A., Meerschaert, M.M., Baeumer, B.: Fractal mobile/immobile solute transport.Water Resour. Res., vol. 39, no. 10, Article ID 1296, 2003. https://doi.org/10.1029/2003WR002141

work page doi:10.1029/2003wr002141 2003

[3] [3]

Elsevier, Amsterdam, 2006

Kilbas, A.A., Srivastava, H.M., Trujillo, J.J.:Theory and Applications of Fractional Differential Equations. Elsevier, Amsterdam, 2006

work page 2006

[4] [4]

Springer, Berlin, 2006

Isakov, V.:Inverse Problems for Partial Differential Equations. Springer, Berlin, 2006

work page 2006

[5] [5]

Theory and Applications

Kabanikhin, S.I.:Inverse and Ill-Posed Problems. Theory and Applications. De Gruyter, 2011

work page 2011

[6] [6]

Sakamoto, K., Yamamoto, M.: Initial value/boundary value problems for fractional diffusion wave equations and applications to some inverse problems.J. Math. Anal. Appl., vol. 382, pp. 426–447, 2011. https://doi.org/10.1016/j.jmaa.2011.04.058

work page doi:10.1016/j.jmaa.2011.04.058 2011

[7] [7]

Lincei Mat

Florida, G., Li, Z., Yamamoto, M.: Well-posedness for the backward problems in time for general time-fractional diffusion equation.Rend. Lincei Mat. Appl., vol. 31, pp. 593– 610, 2020. https://doi.org/10.4171/RLM/906

work page doi:10.4171/rlm/906 2020

[8] [8]

Alimov, Sh., Ashurov, R.: On the backward problems in time for time-fractional subdiffusion equations.Fract. Differ. Calc., vol. 11, no. 2, pp. 203–217, 2021. https://doi.org/10.7153/fdc-2021-11-14 THE BACKW ARD PROBLEM FOR A MULTI-TERM . . . 13

work page doi:10.7153/fdc-2021-11-14 2021

[9] [9]

Wei, T., Xian, J.: Variational method for a backward problem for a time- fractional diffusion equation.ESAIM: M2AN, vol. 53, no. 4, pp. 1223–1244, 2019. https://doi.org/10.1051/m2an/2019019

work page doi:10.1051/m2an/2019019 2019

[10] [10]

Tuan, N.H., Le, H.N., Ngoc, T.B., Zhou, Y.: On a backward problem for non- linear fractional diffusion equations.Appl. Math. Lett., vol. 92, pp. 76–84, 2019. https://doi.org/10.1016/j.aml.2018.11.015

work page doi:10.1016/j.aml.2018.11.015 2019

[11] [11]

Liu, J.J., Yamamoto, M.: A backward problem for the time-fractional diffusion equation. Appl. Anal., vol. 89, pp. 1769–1788, 2010. https://doi.org/10.1080/00036810903479731

work page doi:10.1080/00036810903479731 2010

[12] [12]

Nonlinear Sci

Wen, J., Wang, Y.P., Wang, Y.X., Wang, Y.Q.: The quasi-reversibility regular- ization method for backward problem of the multi-term time-space fractional dif- fusion equation.Commun. Nonlinear Sci. Numer. Simul., Article ID 107848, 2024. https://doi.org/10.1016/j.cnsns.2024.107848

work page doi:10.1016/j.cnsns.2024.107848 2024

[13] [13]

Math., vol

Sun, L., Wang, Y., Chang, M.: A fractional-order quasi-reversibility method to a back- ward problem for the multi-term time-fractional diffusion equation.Taiwanese J. Math., vol. 27, no. 6, pp. 1185–1210, 2023. https://doi.org/10.11650/tjm/230801

work page doi:10.11650/tjm/230801 2023

[14] [14]

Li, Z., Liu, Y., Yamamoto, M.: Initial-boundary value problems for multi-term time- fractional diffusion equations with positive constant coefficients.Appl. Math. Comput., vol. 257, pp. 381–397, 2015. https://doi.org/10.1016/j.amc.2014.11.073

work page doi:10.1016/j.amc.2014.11.073 2015

[15] [15]

https://doi.org/10.48550/arXiv.2603.01833

Ashurov, R., Shamuratov, D.: Inverse problem for a multi-term time- fractional diffusion equation with the Caputo derivatives.arXiv:2603.01833, 2026. https://doi.org/10.48550/arXiv.2603.01833

work page doi:10.48550/arxiv.2603.01833 2026

[16] [16]

Springer, Berlin, 2014

Gorenflo, R., Kilbas, A.A., Mainardi, F., Rogosin, S.:Mittag-Leffler Functions, Related Topics and Applications. Springer, Berlin, 2014. Ravshan Ashurov: V. I. Romanovskiy Institute of Mathematics Uzbekistan Acad- emy of Sciences, University street 9, Tashkent–100174, Uzbekistan Email address:ashurovr@gmail.com Damir Shamuratov: V. I. Romanovskiy Institut...

work page 2014