REVIEW 4 major objections 5 minor 3 cited by
Well-Posedness for the Rosenzweig-MacArthur Model with Internal Stochasticity
T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read This paper claims that a stochastic Rosenzweig-MacArthur predator-prey model with internal demographic noise has a unique strong solution, and that the positive quadrant is invariant, so predator and prey densities remain positive for all…
desk verdict The stochastic Rosenzweig-MacArthur model is a natural new object and the deterministic review is clean, but the central domain-invariance proof collapses on a false Lipschitz-extension step and a Lyapunov function blind to the zero boundary; the main claim is not established and likely false. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the Lyapunov function $V(N,P)=(1+N^2+P^2)^\alpha$ with $\alpha>2$, together with the localization criterion of Theorem 4.3. The criterion says that if the drift $\mu$ and diffusion $\sigma$ are Lipschitz on each truncating set $K_n$, and if there is a nonnegative $V$ with $LV\le \alpha V$ and $\inf_{D\setminus K_n}V\to\infty$, then a unique strong solution exists and the open domain $D$ is invariant. The noise coefficients $\sigma_{11}=\sqrt{N(1+N/k)+mNP/(1+N)}$ and $\sigma_{22}=\sqrt{cP+mNP/(1+N)}$ encode the internal-demographic-noise assumption that variance scales with population density, and it is this structure that makes the generator estimate $LV\le C_1V$ possible. The Euler-Maruyama scheme is then used in simulations to compare stochastic paths with the deterministic limit cycle or stable equilibrium.
What would settle it
Compute the quantity $\inf_{(N,P)\in D\setminus K_n} V(N,P)$ for $V=(1+N^2+P^2)^\alpha$ and $K_n=(0,n]^2$; since $D\setminus K_n$ contains points such as $(\varepsilon,n)$ with $\varepsilon\downarrow0$, the infimum stays near $1+n^2$ rather than going to infinity, so condition (ii) of Theorem 4.3 is not satisfied by the stated $V$. A direct check of whether the proof can be repaired, together with a pathwise simulation that stops at the first hit of zero, would settle whether the claimed invariance of $D$ actually holds.
Extended reading notes
Core claim
The paper's central discovery is Theorem 4.5: for initial data in $D=\{(N,P):N>0,P>0\}$, the stochastic Rosenzweig-MacArthur system (1.4) admits a unique strong solution, and $D$ is invariant, meaning $P((N(t),P(t))\in D)=1$ for every $t\ge 0$. The proof rests on a new criterion (Theorem 4.3) for existence and uniqueness of autonomous SDEs on an open submanifold of $\mathbb{R}^d$, which relaxes earlier localization requirements and permits Lyapunov functions such as $V=(1+N^2+P^2)^\alpha$ with $\alpha>2$. The diffusion coefficients are chosen so that the infinitesimal generator satisfies $LV\le C_1V$, and the criterion then yields both global existence and invariance of the open domain. The paper also derives $p$-th moment estimates and an almost-sure Lyapunov-exponent bound $\limsup_{t\to\infty}\frac{1}{t}\log\|X_t\|\le C_1=\frac{5+6m+c}{4}+\frac{1}{2k}$, which it interprets as at-most-exponential population growth. The invariance of $D$ is emphasized as the rigorous content of biological nonnegativity and is used to conclude that internal stochasticity alone cannot cause extinction.
Load-bearing premise
The localization proof assumes that the Lyapunov function becomes large outside each truncating set, but the chosen function $V=(1+N^2+P^2)^\alpha$ grows only when densities become large, not when a density approaches zero, so the proof does not actually rule out hitting the zero boundary.
Editorial extensions
If this is right
- Starting from any positive initial densities, the solution stays in the positive quadrant for every finite time, so the model predicts that internal demographic fluctuations alone do not produce finite-time extinction.
- The generalized existence-uniqueness criterion applies to other Lotka-Volterra-type systems posed on open submanifolds and permits Lyapunov functions that earlier localization criteria excluded.
- The $p$-th moment and Lyapunov exponent estimates bound population growth at most exponentially, with explicit constants in terms of the model parameters.
- For parameter regimes where the deterministic system has a stable limit cycle or a stable positive equilibrium, the stochastic paths fluctuate around those attractors while remaining positive.
Reading between the lines
- The theorem proves positivity for each fixed finite $t$; it does not by itself prevent $\liminf_{t\to\infty}N(t)$ or $\liminf_{t\to\infty}P(t)$ from being zero, so the paper's language about persistence goes a step beyond what the invariance theorem alone establishes.
- A natural test of the claimed invariance is to classify the boundary behavior of the square-root diffusion coefficients: since the noise vanishes as $N$ or $P$ approaches zero, Feller or boundary-classification criteria could decide whether hitting the boundary in finite time is truly impossible or merely has probability zero under the constructed solution.
- The same Lyapunov-localization template could be used to prove well-posedness for other predator-prey or epidemiological SDEs with state-dependent internal noise and a natural positive-invariant domain.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a stochastic version of the Rosenzweig-MacArthur predator-prey model driven by internal demographic noise, with diffusion coefficients proportional to population densities. It presents a general existence-and-uniqueness criterion for autonomous SDEs on an open submanifold of R^n (Theorem 4.3), then applies it to the stochastic Rosenzweig-MacArthur model to claim global well-posedness and invariance of the positive quadrant D (Theorem 4.5). The paper further derives L^p moment estimates and an almost-sure exponential growth bound, and supports the analysis with Euler-Maruyama simulations showing deviations from the deterministic model.
Significance. If the central results were correct, the paper would offer a useful framework for SDEs constrained to open submanifolds and a rigorous persistence statement for an intrinsically stochastic predator-prey model. The Lyapunov computations in Theorem 4.5 and the L^p estimates in Proposition 4.8 are explicit and checkable, and the general idea of using Lyapunov functions that do not blow up at every boundary is worth exploring. However, the main domain-invariance claim is false: the chosen Lyapunov function does not blow up at the zero boundary, and the square-root noise structure makes that boundary attainable. The proof of the general criterion also contains a load-bearing gap. Since the paper's biological conclusion of persistence and non-extinction rests on the invalid Theorem 4.5, the current manuscript is not suitable for publication.
major comments (4)
- [Section 4, Theorem 4.5] The sets K_n = (0,n]^2 are closed in D but not precompact: their closure in D is still (0,n]^2, which is not compact because a sequence such as (1/m,1) has no convergent subsequence in D. Hence no open precompact subset ~K_n of D can contain K_n, and the hypothesis of Theorem 4.3 requiring the existence of such ~K_n fails. Therefore Theorem 4.3 cannot be applied to the model, and the proof of Theorem 4.5 collapses at this point.
- [Section 4, Theorem 4.3 proof] The assertion that 'since for each n, supp ~mu_n, supp ~sigma_n ⊆ ~K_n and ~K_n is precompact, it is easy to see each mu_n and sigma_n are globally Lipschitz continuous' is false: a continuous function with compact support need not be Lipschitz (e.g., sqrt(x) on [0,1]). In the application the coefficients are C^1 on D, so a smooth extension with compact support could be used instead, but the argument as written does not establish the Lipschitz condition required for the classical existence theorem.
- [Section 4, Theorem 4.5 proof and Remark 4.6] The Lyapunov function V = (1+N^2+P^2)^alpha is bounded on the strips 0<N<=n, 0<P<=n near the axes. Consequently the estimate P(tau_n <= t) <= ... / inf_{D\K_n} V only controls the event that the process escapes to large values of N or P; it gives no control over the event that N or P hits zero. If the process exits D through the zero boundary, then X_{tau_n} is not in D, V is not defined at that point, and the lower bound V(X_{tau_n}) >= inf_{D\K_n} V used in the proof fails. Thus the proof does not establish P((N(t),P(t)) in D for all t) = 1.
- [Section 4, Theorem 4.5 statement] The claimed invariance is not merely unproved but is false. Near N=0 with P>0, the prey equation in (1.4) behaves like dN ≈ (1-mP)N dt + sqrt(N(1+mP)) dB^1. Writing Y = sqrt(N) gives dY ≈ [(1-mP)Y/2 - (1+mP)/(8Y)] dt + (1/2) sqrt(1+mP) dB^1, whose -1/Y drift is that of a Bessel-type process with dimension less than 2. Feller's boundary classification then shows that 0 is an attainable boundary, so with positive probability N can hit zero in finite time. Hence P((N(t),P(t)) in D for all t>=0) = 1 cannot hold for general initial data, contradicting Theorem 4.5 and the persistence interpretation in Remark 4.6.
minor comments (5)
- [Section 4, Lemma 4.1] In the statement of Lemma 4.1, the equation is labelled (3.2), but it should be (4.2) to match the surrounding text.
- [Section 4, Theorem 4.3] In the definition of the infinitesimal generator, the domain of f should be R+×U, not R+×R+, since V is only defined on U.
- [Section 5, Figure 2] The figure caption is internally inconsistent: it refers to panels (a), (b), (e) for the source case and (c), (d), (f) for the sink case, but then describes the trajectory as panel (c) and (f), while the text also mentions panel (e). The labels should be corrected.
- [Throughout] There are several typographical errors, including 'wiith' in Lemma 4.2 and 'mainfolds' in Section 3; a careful proofreading pass is needed.
- [References] The stability and limit-cycle results for the deterministic Rosenzweig-MacArthur model are cited to an unpublished note [Gun12]; the authors should cite the original literature (e.g., Rosenzweig and MacArthur 1963, and Cheng 1981) for these classical results.
Circularity Check
No circularity: the proof chain is an explicit Lyapunov/localization computation, and the zero-boundary gap is a mathematical correctness issue, not a reduction to the paper's own inputs.
full rationale
The paper's claim chain is: Theorem 4.3 gives an existence/uniqueness/domain-invariance criterion via localization and a Lyapunov function; Theorem 4.5 applies that criterion to (1.4) with V=(1+N^2+P^2)^α and verifies LV≤C1V by explicit term-by-term inequalities. No parameter is fitted and then relabelled as a prediction, no result is defined in terms of the target conclusion, and no load-bearing assertion rests on a self-citation by the present authors. The cited results ([Abu91], [Kha12], [Mao07], [Fri75]) are standard external theorems, and Theorem 4.3 is also proved in the paper rather than merely imported. A genuine mathematical flaw exists: K_n=(0,n]^2 is not closed, and V remains bounded near N=0 or P=0, so the localization argument does not exclude hitting the zero boundary; the square-root diffusion near zero resembles a Bessel-type process for which zero can be attainable. But this is a failure of the application of the Lyapunov criterion, not a circular derivation. The invariance conclusion is not equivalent to the assumptions by construction, and there is no fitted-input-predicted-output reduction. Accordingly, the circularity score is 0.
Assumptions & free parameters
assumptions (4)
- ad hoc to paper The truncated extensions ~mu_n and ~sigma_n are globally Lipschitz continuous
- domain assumption Each truncating set K_n has an open precompact neighborhood in U
- standard math Classical strong existence and uniqueness for globally Lipschitz SDEs
- domain assumption The diffusion terms in (1.4) keep the process inside D
Cite this review
Pith. "Pith review of Well-Posedness for the Rosenzweig-MacArthur Model with Internal Stochasticity." pith.science (2026). https://pith.science/paper/ORLT6FUG
@misc{pith2026250516904,
author = {Pith},
title = {Pith review of: Well-Posedness for the Rosenzweig-MacArthur Model with Internal Stochasticity},
year = {2026},
howpublished = {\url{https://pith.science/paper/ORLT6FUG}},
note = {Machine review of arXiv:2505.16904}
}
abstract
In this work, we propose a stochastic version of the Rosenzweig-MacArthur model solely driven by internal demographic noise, extending classical Lotka-Volterra-type systems focused on external noise. We give a criterion for the existence and uniqueness of autonomous stochastic differential equations (SDEs) on an open submanifold of $\mathbb{R}^{n}$, and the framework allows for a wider choice of Lyapunov functions. In the meantime, the invariance of open submanifolds, which is a biologically feasible result and has been implicitly incorporated into many biological and ecological models, facilitates the application of analytic tools typically suited to $\mathbb{R}^{d}$ and indicates the persistence of predator and prey populations, thus providing a criterion for determining whether a population will become extinct. We apply the well-posedness criterion to our stochastic Rosenzweig-MacArthur model and show the existence and uniqueness of solutions. Furthermore, the asymptotic estimates of solutions are obtained, indicating the at most exponential growth of the population with internal stochasticity. Some numerical experiments are performed, which illustrate the discrepancy between the deterministic and stochastic models. Overall, this work demonstrates the broad applicability of our results to ecological models with constrained dynamics, offering a foundation for analyzing extinction, persistence, and well-posedness in systems where internal randomness dominates. This paper not only promotes the development of stochastic modeling and stochastic differential equations in theoretical ecology but also proposes a rigorous mathematical methodology for studying the predator-prey system with internal stochasticity.
Forward citations
Cited by 3 Pith papers
-
Reflected Optimal Stopping with a Max-Type Payoff: Measure-Valued Stopping Gains and Killed Resolvent Representation
For reflected diffusion optimal stopping with G=x1∨αx2, the gain Γ is a signed measure with singular diagonal part −(n⊤an)/(2√(1+α²))σΔ, and V=G−R_C^r Γ (killed at first entry to the stopping set).
-
Multiplier Sensitivity in Isoperimetric Optimal Control
Under stated convexity and attainability conditions, the isoperimetric Pontryagin multiplier is the negative supergradient of the constrained value function, and this sensitivity can survive when the modified control-...
-
A Full-Density Approach to Simulating Random Iteration Equations with Applications
A full-density propagation method simulates random iteration equations without repeated Monte Carlo sampling and is applied to stochastic DEs, uncertainty-aware optimization, and chaotic maps.
Reference graph
Works this paper leans on
-
[1]
A stochastic model for predator-prey systems: basic properties, stability and computer simulation
Mr Abundo. A stochastic model for predator-prey systems: basic properties, stability and computer simulation. Journal of Mathematical Biology , 29:495--511, 1991
work page 1991
-
[2]
Ludwig Arnold and Ren \'e Lefever. Stochastic Nonlinear Systems in Physics, Chemistry, and Biology: Proceedings of the Workshop Bielefeld, Fed. Rep. of Germany, October 5--11, 1980 , volume 8. Springer Science & Business Media, 2012
work page 1980
-
[3]
Uniqueness of a limit cycle for a predator-prey system
Kuo-Shung Cheng. Uniqueness of a limit cycle for a predator-prey system. SIAM Journal on Mathematical Analysis , 12(4):541--548, 1981
work page 1981
-
[4]
An introduction to population genetics theory
James Franklin Crow. An introduction to population genetics theory . Scientific Publishers, 2017
work page 2017
-
[5]
Marcus W Feldman and Jonathan Roughgarden. A population's stationary distribution and chance of extinction in a stochastic environment with remarks on the theory of species packing. Theoretical population biology , 7(2):197--207, 1975
work page 1975
-
[6]
Stochastic differential equations and applications
Avner Friedman. Stochastic differential equations and applications. In Stochastic differential equations , pages 75--148. Springer, 1975
work page 1975
-
[7]
On the volterra and other nonlinear models of interacting populations
Narendra S Goel, Samaresh C Maitra, and Elliott W Montroll. On the volterra and other nonlinear models of interacting populations. Reviews of modern physics , 43(2):231, 1971
work page 1971
-
[8]
The rosenzweig-macarthur predator-prey model
Chathika Gunaratne. The rosenzweig-macarthur predator-prey model. https://www.academia.edu/6247163/THE_ROSENZWEIG_MACARTHUR_PREDATOR_PREY_MODEL, 2012. Accessed: 2025-05-21
Show all 22 references
-
[9]
Nonlinear dynamical systems and control: a Lyapunov-based approach
Wassim M Haddad and VijaySekhar Chellaboina. Nonlinear dynamical systems and control: a Lyapunov-based approach . Princeton university press, 2008
2008
-
[10]
Stochastic stability of differential equations
Rafail Khasminskii. Stochastic stability of differential equations . Springer, 2012
2012
-
[11]
Stochastic differential equations
Peter E Kloeden, Eckhard Platen, Peter E Kloeden, and Eckhard Platen. Stochastic differential equations . Springer, 1992
1992
-
[12]
Brownian motion and stochastic calculus , volume 113
Ioannis Karatzas and Steven Shreve. Brownian motion and stochastic calculus , volume 113. Springer Science & Business Media, 1991
1991
-
[13]
Probl \`e me g \'e n \'e ral de la stabilit \'e du mouvement
Alexandre Liapounoff. Probl \`e me g \'e n \'e ral de la stabilit \'e du mouvement. Annales de la Facult \'e des sciences de l'Universit \'e de Toulouse pour les sciences math \'e matiques et les sciences physiques , 9:203--474, 1907
1907
-
[14]
Smooth manifolds
John M Lee and John M Lee. Smooth manifolds . Springer, 2012
2012
-
[15]
Persistence of dynamical systems under random perturbations
Donald Ludwig. Persistence of dynamical systems under random perturbations. Siam Review , 17(4):605--640, 1975
1975
-
[16]
The general problem of the stability of motion
Aleksandr Mikhailovich Lyapunov. The general problem of the stability of motion. International journal of control , 55(3):531--534, 1992
1992
-
[17]
Stochastic differential equations and applications
Xuerong Mao. Stochastic differential equations and applications . Elsevier, 2007
2007
-
[18]
Stability and complexity in model ecosystems
Robert M May. Stability and complexity in model ecosystems . Princeton university press, 2019
2019
-
[19]
Mathematical biology: I
James D Murray. Mathematical biology: I. An introduction , volume 17. Springer Science & Business Media, 2007
2007
-
[20]
Differential equations and dynamical systems , volume 7
Lawrence Perko. Differential equations and dynamical systems , volume 7. Springer Science & Business Media, 2013
2013
-
[21]
The deterministic theory of population dynamics
Aldo Rescigno and IW Richardson. The deterministic theory of population dynamics. In Foundations of mathematical biology , pages 283--360. Elsevier, 1973
1973
-
[22]
write newline
" write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTION or pop #1 'skip if FUNCTION new.block.check...
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.