arxiv: 2508.17731 · v4 · submitted 2025-08-25 · 🧮 math.OC · math.PR

G-BSDEs with non-Lipschitz coefficients and the corresponding stochastic recursive optimal control problem

Wei He , Qiangjun Tang This is my paper

Pith reviewed 2026-05-18 21:53 UTC · model grok-4.3

classification 🧮 math.OC math.PR

keywords G-BSDEnon-Lipschitz coefficientsstochastic recursive controldynamic programming principleviscosity solutionHamilton-Jacobi-Bellman equationEpstein-Zin utility

0 comments

The pith

G-BSDEs with non-Lipschitz generators admit unique solutions under uniform continuity and monotonicity in one variable.

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

The paper proves existence and uniqueness for G-BSDEs whose generators are uniformly continuous and monotonic with respect to the first variable but need not be Lipschitz. It applies the results to a stochastic recursive optimal control problem, establishing the dynamic programming principle and showing that the value function is the viscosity solution of the associated Hamilton-Jacobi-Bellman equation. A reader would care because the conditions cover generators arising in recursive utility models, such as continuous-time Epstein-Zin preferences, that fall outside standard Lipschitz theory.

Core claim

Under the assumption that the generator of the G-BSDE is uniformly continuous and monotonic with respect to the first unknown variable, the equation admits a unique solution. For the corresponding stochastic recursive optimal control problem, the value function satisfies the dynamic programming principle and is the viscosity solution of the Hamilton-Jacobi-Bellman equation.

What carries the argument

The comparison theorem for G-BSDEs combined with stability of viscosity solutions

If this is right

The value function of the recursive control problem is characterized as the viscosity solution of the HJB equation.
The dynamic programming principle holds for the stochastic recursive optimal control problem.
The framework covers continuous-time Epstein-Zin utility as a special case.
Existence and uniqueness results apply to G-BSDEs beyond the Lipschitz setting.

Where Pith is reading between the lines

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

The results widen the class of admissible generators in recursive utility maximization.
Comparable arguments may extend to other nonlinear expectation frameworks.
The HJB characterization opens a route to numerical solution methods for the control problem.

Load-bearing premise

The generator must be uniformly continuous and monotonic with respect to its first argument.

What would settle it

A concrete G-BSDE whose generator is monotonic and continuous but fails uniform continuity, for which either existence or uniqueness of solutions does not hold.

read the original abstract

In this paper, we study the existence and uniqueness of solutions to a class of non-Lipschitz G-BSDEs and the corresponding stochastic recursive optimal control problem. More precisely, we suppose that the generator of G-BSDE is uniformly continuous and monotonic with respect to the first unknown variable. Using the comparison theorem for G-BSDE and the stability of viscosity solutions, we establish the dynamic programming principle and the connection between the value function and the viscosity solution of the associated Hamilton-Jacobi-Bellman equation.We provide an example of continuous time Epstein-Zin utility to demonstrate the application of our study.

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

They prove existence and uniqueness for G-BSDEs under uniform continuity plus monotonicity in y, then use comparison and viscosity stability to get the DPP and value-function/HJB link, with an Epstein-Zin example.

read the letter

The one thing to know is that the authors drop the Lipschitz condition on the G-BSDE generator down to uniform continuity plus monotonicity in the first variable, prove existence and uniqueness, and then show how this lets them establish the dynamic programming principle for the stochastic recursive optimal control problem along with the viscosity solution property for the value function. The Epstein-Zin utility example illustrates the point for ambiguity settings. What the paper does well is take established comparison and viscosity techniques and apply them in this weaker regularity class without overclaiming a big methodological shift. The example is helpful because it points to a concrete use case in recursive preferences where the non-Lipschitz generator might arise naturally. The soft spot is around whether the comparison theorem holds under those exact assumptions without additional hidden conditions. The manuscript appears to state the assumptions up front and then invoke the theorem after proving the BSDE existence, so there is no obvious circularity. But the non-Lipschitz case often requires extra care in the a priori estimates, and that is the part that would need close checking in a review. This paper is for people working on G-stochastic control or nonlinear expectations who want to handle generators with less regularity. A reader who follows the G-BSDE literature or works on ambiguity in continuous-time control would get something out of the extension and the example. It is a targeted technical result rather than a broad new framework. I would send it to peer review. The problem is well-posed, the assumptions are explicit, and the high-level argument is coherent.

Referee Report

2 major / 3 minor

Summary. The manuscript studies existence and uniqueness for G-BSDEs whose generators are uniformly continuous and monotone in the y-variable (rather than Lipschitz). It then treats the associated stochastic recursive optimal control problem, invoking the comparison theorem for G-BSDEs together with viscosity-solution stability to obtain the dynamic programming principle and to identify the value function as a viscosity solution of the corresponding HJB equation. The results are illustrated by a continuous-time Epstein-Zin utility example.

Significance. If the derivations hold, the work extends G-BSDE theory and recursive control beyond the Lipschitz regime, which is relevant for applications such as Epstein-Zin preferences where the Lipschitz condition is typically violated. The explicit use of comparison and viscosity stability to reach the DPP and HJB link supplies a concrete route for handling these problems.

major comments (2)

[§3] §3 (existence/uniqueness for the G-BSDE): the uniqueness argument relies on the monotonicity assumption, but the estimates appear to require a quantitative modulus from uniform continuity; it is not immediate that the same contraction or Gronwall-type bound closes without an implicit Lipschitz constant. A short explicit estimate or counter-example check would strengthen the claim.
[§4.2] §4.2 (dynamic programming principle): the passage from the comparison theorem to the DPP for the controlled G-BSDE requires that the monotonicity and uniform continuity hold uniformly with respect to the control parameter; the manuscript should verify that the generator family indexed by admissible controls satisfies the hypotheses of the invoked comparison result.

minor comments (3)

[§2] The statement of the generator assumptions (uniform continuity plus monotonicity in y) should be collected in a single numbered hypothesis at the beginning of §2 for easy reference.
[§5] In the Epstein-Zin example, the explicit form of the generator and the verification that it meets the uniform-continuity/monotonicity conditions should be written out rather than left as a reference to the literature.
[Introduction] A few sentences comparing the present non-Lipschitz setting with earlier Lipschitz results for G-BSDEs would help readers gauge the precise advance.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comments. We address the two major comments point by point below and have revised the paper to improve clarity where needed.

read point-by-point responses

Referee: [§3] §3 (existence/uniqueness for the G-BSDE): the uniqueness argument relies on the monotonicity assumption, but the estimates appear to require a quantitative modulus from uniform continuity; it is not immediate that the same contraction or Gronwall-type bound closes without an implicit Lipschitz constant. A short explicit estimate or counter-example check would strengthen the claim.

Authors: We thank the referee for highlighting this point. The uniqueness proof in Section 3 combines the monotonicity condition (which produces a negative drift term for the y-difference) with the uniform continuity of the generator (which supplies a modulus of continuity ω(·) to bound the generator difference). After localization via stopping times, the resulting inequality is closed by a Gronwall-type argument that incorporates the modulus ω directly, without requiring a global Lipschitz constant. To address the concern, we have added an explicit step-by-step estimate in the revised Section 3 that displays how the modulus enters the bound and confirms the contraction closes. revision: yes
Referee: [§4.2] §4.2 (dynamic programming principle): the passage from the comparison theorem to the DPP for the controlled G-BSDE requires that the monotonicity and uniform continuity hold uniformly with respect to the control parameter; the manuscript should verify that the generator family indexed by admissible controls satisfies the hypotheses of the invoked comparison result.

Authors: We agree that uniformity with respect to the control must be verified. In the problem formulation, the generator takes the form f(t, y, z, α) with α ranging over a compact control set A, and both the uniform continuity modulus and the monotonicity constant are assumed independent of α (i.e., they hold uniformly over A). Consequently, every generator in the family indexed by admissible controls satisfies the hypotheses of the comparison theorem uniformly. We have inserted a short verification paragraph in the revised Section 4.2 that explicitly records this uniformity and justifies direct application of the comparison result. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper proves existence and uniqueness for non-Lipschitz G-BSDEs under explicit assumptions of uniform continuity and monotonicity in the generator. It then invokes the comparison theorem for G-BSDEs and stability of viscosity solutions as external standard results to obtain the dynamic programming principle and the value-function/HJB connection. No derivation step reduces by construction to a fitted input, self-definition, or load-bearing self-citation chain; the central claims rest on independent prior theorems in the field rather than internal circular reductions.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that the generator satisfies uniform continuity and monotonicity in the first variable; no free parameters or new invented entities are introduced in the abstract.

axioms (1)

domain assumption The generator of the G-BSDE is uniformly continuous and monotonic with respect to the first unknown variable
Explicitly stated as the supposition under which existence, uniqueness, and the dynamic programming principle are proved.

pith-pipeline@v0.9.0 · 5625 in / 1309 out tokens · 45108 ms · 2026-05-18T21:53:03.255307+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

we suppose that the generator of G-BSDE is uniformly continuous and monotonic with respect to the first unknown variable. Using the comparison theorem for G-BSDE and the stability of viscosity solutions, we establish the dynamic programming principle
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

G-Brownian motion; Stochastic recursive optimal control; Backward SDEs; Non-Lipschitz

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages · 1 internal anchor

[1]

Briand, Ph., Delyon, B., Hu, Y., Pardoux, E., Stoica, L.: Lp solutions of backward stochastic differential equations. Stoch. Process. Appl. 108(1), 109-129 (2003) 31

work page 2003
[2]

Crandall, M.G., Ishii, H., Lions, P.L.: User’s guide to viscosity solutions of second order partial differential equations. Bull. Am. Math. Soc. 27(1), 1-67 (1992)

work page 1992
[3]

Econometrica 60(2), 353-394 (1992)

Duffie, D., Epstein, L.: Stochastic differential utility. Econometrica 60(2), 353-394 (1992)

work page 1992
[4]

Potential Anal

Denis, L., Hu, M., Peng S.: Function spaces and capacity related to a sublinear expectation: application to G-Brownian motion pathes. Potential Anal. 34(2), 139-161 (2011)

work page 2011
[5]

Springer Verlag (2006)

Fleming, W.H., Soner, H.M.: Controlled Markov Processes and Viscosity Solutions. Springer Verlag (2006)

work page 2006
[6]

He, W.: BSDEs driven by G-Brownian motion with non-Lipschitz coefficients. J. Math. Anal. Appl. 505(2) 125569 (2022)

work page 2022
[7]

Hu, M., Ji, S.: Dynamic programming principle for stochastic recursive optimal control problem driven by a G-Brownian motion. Stoch. Process. Appl. 127(1), 107-134 (2017)

work page 2017
[8]

Hu, M., Ji, S., Peng, S., Song, Y.: Backward stochastic differential equations driven by G- Brownian motion. Stoch. Process. Appl. 124(1), 759-784 (2014)

work page 2014
[9]

Hu, M., Ji, S., Peng, S., Song, Y.:Comparison theorem, Feynman-Kac formula and Girsanov transformation for BSDEs driven by G-Brownian motion. Stoch. Process. Appl. 124(2), 1170- 1195 (2014)

work page 2014
[10]

Hu, M., Qu, B., Wang, F.: BSDEs driven by G-Brownian motion with time-varying Lipschitz condition. J. Math. Anal. Appl. 491(2), 124342 (2020)

work page 2020
[11]

ESAIM Control Optim

Hu, M., Wang, F.: Stochastic optimal control problem with infinite horizon driven byG-Brownian motion. ESAIM Control Optim. Calc. Var. 24(2), 873-899 (2018)

work page 2018
[12]

ESAIM Control Optim

Hu, M., Ji, S., Li, X.: Dynamic programming principle and Hamilton-Jacobi-Bellman equation under nonlinear expectation. ESAIM Control Optim. Calc. Var. 28(5), 1-21 (2022)

work page 2022
[13]

Hu, M., Wang, F., Zheng, G.: Quasi-continuous random variables and processes under the G expectation framework. Stoch. Process. Appl. 126(8), 2367-2387(2016)

work page 2016
[14]

Hu, Y., Peng, S.: Solution of forward-backward stochastic differential equations. Probab. Theory Relat. Fields 103(2), 273-283 (1995)

work page 1995
[15]

Springer (2014)

Karatzas, I., Shreve, S.: Brownian Motion and Stochastic Calculus. Springer (2014)

work page 2014
[16]

Finance Stoch

Kraft, H., Seifried, F.T., Steffensen, M.: Consumption-portfolio optimization with recursive util- ity in incomplete markets. Finance Stoch. 17(1), 161-196 (2013)

work page 2013
[17]

Israel J

Lasry, J.M., Lyons, P.L.: A remark on regularization in Hilbert spaces. Israel J. Math. 55(3), 257-266 (1986)

work page 1986
[18]

Lepeltier, J.P., San Martin, J.: Backward stochastic differential equations with continuous coef- ficient. Stat. Probab. Lett. 32(5), 425-430 (1997)

work page 1997
[19]

Nonlinear Anal.70(4), 1776-1796 (2009)

Li, J., Peng, S.: Stochastic optimization theory of backward stochastic differential equations with jumps and viscosity solutions of Hamilton-Jacobi-Bellman equations. Nonlinear Anal.70(4), 1776-1796 (2009)

work page 2009
[20]

Stochastic 92(5), 659-683 (2020) 32

Liu, G.: Multi-dimensional BSDEs driven by G-Brownian motion and related system of fully nonlinear PDEs. Stochastic 92(5), 659-683 (2020) 32

work page 2020
[21]

arXiv preprint arXiv:2104.06868 (2021)

Lu, H., Song, Y.: Forward-backward stochastic differential equations driven by G-Brownian motion. arXiv preprint arXiv:2104.06868 (2021)

work page arXiv 2021
[22]

Kluwer Academic Publishers, Dordrecht, 503-549 (1999)

Pardoux, E.: BSDE’s, weak convergence and homogenization of semilinear PDE’s, in Nonlinear Analysis, Differential Equations and Control. Kluwer Academic Publishers, Dordrecht, 503-549 (1999)

work page 1999
[23]

Pardoux, E., Tang, S.: Forward-backward stochastic differential equations and quasilinear parabolic PDEs. Probab. Theory Relat. Fields 114(2), 123-150 (1999)

work page 1999
[24]

Stochastics 37(1-2), 61-74 (1991)

Peng, S.: Probabilistic Interpretation for Systems of Quasilinear Parabolic Partial Differential Equations. Stochastics 37(1-2), 61-74 (1991)

work page 1991
[25]

Stochastics 38, 119-134 (1992)

Peng, S.: A Generalized Dynamic Programming Principle and Hamilton-Jacobi-Bellmen equa- tion. Stochastics 38, 119-134 (1992)

work page 1992
[26]

In Topics on Stochastic Analysis, edited by Yan, J., Peng, S., Fang, S., Wu, L

Peng, S.: Backward stochastic differential equations-stochastic optimization theory and viscosity solutions of HJB equations. In Topics on Stochastic Analysis, edited by Yan, J., Peng, S., Fang, S., Wu, L. Science Press, Beijing (in Chinese), 85-138 (1997)

work page 1997
[27]

Stochastic Analysis and Applications, 541-567, Abel Symposium, 2, Springer, Berlin (2007)

Peng, S.: G-expectation, G-Brownian Motion and Related Stochastic Calculus of Itˆ o type. Stochastic Analysis and Applications, 541-567, Abel Symposium, 2, Springer, Berlin (2007)

work page 2007
[28]

Nonlinear Expectations and Stochastic Calculus under Uncertainty

Peng, S.: Nonlinear Expectations and Stochastic Calculus under Uncertainty. arXiv preprint arXiv:1002.4546v1 (2010)

work page internal anchor Pith review Pith/arXiv arXiv 2010
[29]

Probability Theory and Stochastic Modelling 95, Springer (2019)

Peng, S.: Nonlinear Expectations and Stochastic Calculus Under Uncertainty: With Robust CLT and G-Brownian Motion. Probability Theory and Stochastic Modelling 95, Springer (2019)

work page 2019
[30]

Possama¨ ı, D.: Second order backward stochastic differential equations under a monotonicity condition. Stoch. Process. Appl. 123(5), 1521-1545 (2013)

work page 2013
[31]

ESAIM Control Optim

Pu, J., Zhang, Q.: Dynamic programming principle and associated Hamilton-Jacobi-Bellman equation for stochastic recursive control problem with non-Lipschitz aggregator. ESAIM Control Optim. Calc. Var. 24(1), 355-376 (2018)

work page 2018
[32]

Soner, M., Touzi, N., Zhang, J.: well-posedness of Second Order Backward SDEs. Probab. Theory Relat. Fields 153(1-2), 149-190 (2012)

work page 2012
[33]

Song, Y.: Some properties on G-evaluation and its applications to G-martingale decomposition. Sci. China Math. 54(2), 287-300 (2011)

work page 2011
[34]

Wang, F., Zheng, G.: Backward stochastic differential equations driven by G-Bsrownian motion with uniformly continuous generators. J. Theor. Probab. 34(2), 660-681 (2021)

work page 2021
[35]

arXiv Preprint arXiv:2205.03052v3 (2022)

Wen, J., Wu, Z., Zhang, Q.: Dynamic programming principle for delayed stochastic recur- sive optimal control problem and HJB equation with non-Lipschitz generator. arXiv Preprint arXiv:2205.03052v3 (2022)

work page arXiv 2022
[36]

Wu, Z., Yu, Z.: Dynamic programming principle for one kind of stochastic recursive optimal control problem and Hamilton-Jacobi-Bellman equation. SIAM J. Control. Optim. 47(5), 2616- 2641 (2008)

work page 2008
[37]

arXiv Preprint arXiv:2409.16574v1 (2024) 33

Zhao, B.: BSDEs driven by G-Brownian motion with time-varying uniformly continuous genera- tors. arXiv Preprint arXiv:2409.16574v1 (2024) 33

work page arXiv 2024
[38]

Zhuo, Y., Dong, Y., Pu, J.: Dynamic programming principle and viscosity solutions of Hamilton- Jacobi-Bellman equations for stochastic recursive control problem with non-Lipschitz generator. Appl. Math. Optim. 82(2), 851-887 (2019) 34

work page 2019