Pith. sign in

REVIEW 3 major objections 5 minor 36 references

Suppression of blow up by mixing in generalized Keller-Segel system with fractional dissipation

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Mixing by a weakly mixing flow suppresses blow-up in the generalized Keller-Segel equation for every fractional diffusion order $0<\alpha<2$ in all dimensions $d\ge2$.

desk verdict Real parameter-range extension with a repairable but real gap in the heart of the proof; worth refereeing with expectation of major revision. read the letter →

arxiv 1908.02489 v1 pith:QNKYTQ3H submitted 2019-08-07 math.AP

Binbin Shi , Weike Wang This is my paper · ORCID
classification math.AP MSC 35A0135B4535R1135Q92
keywords Keller-SegelequationsfractionaldissipationmixingrelaxationenhancingflowsRAGEtheoremnonlinearmaximumprincipleblow-upsuppressionglobalclassicalsolution
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 finite-time blow-up, known to occur for large-data solutions of the unadvected generalized Keller-Segel equation with fractional dissipation, is suppressed completely when a suitable incompressible flow is added. For every fractional diffusion order $0<\alpha<2$, every $\beta\in[2,d]$, and every dimension $d\ge2$, the authors construct a smooth, weakly mixing divergence-free flow $u$ such that the advected equation (1.1) has a unique global-in-time classical solution for every non-negative initial datum in $H^3(\mathbb{T}^d)\cap L^\infty(\mathbb{T}^d)$. This covers parameter ranges inaccessible to earlier suppression results, which required stronger dissipation such as $\alpha>d/2$ or $\alpha=2$. The proof couples a nonlinear maximum principle on the torus with the relaxation-enhancing effect of weak mixing, which forces the $L^2$ deviation of the density from its mean to contract over short time substeps.

What carries the argument

Four components carry the argument. (i) The weakly mixing incompressible flow $u$, a divergence-free vector field whose transport group $U^t$ has purely continuous spectrum; this is the mixing agent. (ii) The RAGE theorem (Lemma 3.5), which states that for any compact set $K$ of unit vectors in $L^2$ and any finite-rank projection $P_N$ onto the first $N$ eigenfunctions of $(-\Delta)^{\alpha/2}$, the long-time average $T^{-1}\int_0^T\|P_N U^t\varphi\|_{L^2}^2\,dt$ is uniformly small over $\varphi\in K$; this gives the quantitative enhancement of dissipation by mixing. (iii) The nonlinear maximum principle on the torus (Lemma 5.1), which at a spatial maximum $x_t$ of $\rho$ yields either $\rho(t,x_t)\le C\|\rho\|_{L^p}$ or the pointwise lower bound $(-\Delta)^{\alpha/2}\rho(t,x_t)\ge C\rho(t,x_t)^{1+p\alpha/d}\|\rho\|_{L^p}^{-p\alpha/d}$, reducing the maximum evolution to a differential inequality. (iv) The $L^\infty$-criterion (Propositions 3.1 and 4.1), which upgrades a uniform $L^\infty$ bound to an $H^3$ bound through a Gagliardo-Nirenberg interpolation energy estimate. The Proposition 3.4 iteration chooses a large coupling constant $A$ so that each substep of duration $\tau=T_c/A$ contracts $\|\rho-\bar\rho\|_{L^2}$ by the fixed factor $1-\lambda_N^{\alpha/2}\tau/2$.

What would settle it

A concrete check is to test whether (3.39) implies $\|(\rho_0-\bar\rho)/\|\rho_0-\bar\rho\|_{L^2}\|_{\dot H^{\alpha/2}}^2\le\lambda_N^{\alpha/2}$; if some admissible initial datum satisfies (3.39) but not this inequality, then (3.43) fails and the contraction (3.49)--(3.51), on which the global $L^\infty$ bound rests, is unsupported. A numerical iteration of the substep map would also reveal whether the normalized $\dot H^{\alpha/2}$ norm stays below $\lambda_N^{\alpha/2}$ at each step.

Watch

Extended reading notes

Core claim

Theorem 1.1 states that for $0<\alpha<2$, $\beta\in[2,d]$, $d\ge2$, and any non-negative $\rho_0\in H^3(\mathbb{T}^d)\cap L^\infty(\mathbb{T}^d)$, there exists a smooth incompressible flow $u$—weakly mixing in the sense that its transport operator has purely continuous spectrum—such that the unique solution $\rho$ of (1.1) is global in time and lies in $C(\mathbb{R}_+;H^3(\mathbb{T}^d))$. The discovery is that mixing supplies the missing dissipation at every scale: the RAGE time-averaging bound shrinks the $L^2$ fluctuation $\|\rho-\bar\rho\|_{L^2}$ over intervals of length $\tau=T_c/A$, and the nonlinear maximum principle converts the resulting $L^p$ control into a uniform-in-time $L^\infty$ bound. Once the $L^\infty$ norm is controlled, the $L^\infty$-criterion upgrades the solution to a global classical solution. The proof treats separately the singular kernel case $\beta=d$, where $\Delta K$ is not integrable and $B(\rho)=\nabla(-\Delta)^{-1}\rho$, and the case $2\le\beta<d$, where $\Delta K\in L^1$.

Load-bearing premise

The proof of Proposition 3.4 requires that at the start of every substep $t_j=j\tau$ the normalized fluctuation $(\rho(t_j)-\bar\rho)/\|\rho(t_j)-\bar\rho\|_{L^2}$ lies in the fixed compact set $K$ defined by $\|\varphi\|_{\dot H^{\alpha/2}}^2\le\lambda_N^{\alpha/2}$, but the paper only asserts this membership for the initial datum and does not show that the spectral cutoff $N$ chosen in (3.39) forces it for later iterates, whose $L^2$ deviations shrink so their normalized $\dot H^{\alpha/2}$ norms can grow.

Editorial extensions

If this is right

  • Large initial data that would blow up in finite time for the unadvected equation become globally well-posed once advected by a suitably chosen weakly mixing flow, for every $0<\alpha<2$, $\beta\in[2,d]$, $d\ge2$.
  • The global solution belongs to $C(\mathbb{R}_+;H^3(\mathbb{T}^d))$, and the same argument upgrades to $C(\mathbb{R}_+;H^k(\mathbb{T}^d))$ whenever the initial datum lies in $H^k$, $k\ge2$.
  • The result includes the supercritical fractional regime $\alpha<d/2$, where the unadvected equation is known to blow up and where no prior mixing-suppression theorem applied.
  • Both the non-integrable kernel case $\beta=d$ and the integrable-kernel case $2\le\beta<d$ are covered by the same strategy, with the energy estimates adapted to the structure of $\Delta K$.
  • The mixing flow can be chosen in advance from the class of weakly mixing flows; no smallness condition on the initial data is imposed.

Reading between the lines

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

  • The same substep contraction mechanism should be reproducible with other dissipative operators—for instance hyperdissipation or nonlinear diffusion—provided a nonlinear maximum principle and an $L^\infty$-criterion are available; the paper's structure suggests mixing suppresses blow-up whenever those two ingredients exist.
  • A numerical implementation of the substep map with a simple weakly mixing flow (e.g., a time-periodic shear) could measure the actual contraction rate of $\|\rho-\bar\rho\|_{L^2}$ and check whether it matches the predicted factor $1-\lambda_N^{\alpha/2}\tau/2$, which would make the RAGE time $T_c$ the operative bottleneck.
  • The authors note that the range $d<\beta<d+1$ remains open because $\Delta K$ is not integrable there; one natural extension is to split the kernel into a singular part controlled by the $H^3$ energy and a regular part handled by the same mixing iteration.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper studies the generalized parabolic-elliptic Keller-Segel system on the torus with fractional dissipation (−Δ)^{α/2}, 0<α<2, β∈[2,d], d≥2, and advection by an incompressible flow. The main result, Theorem 1.1, asserts that for every nonnegative initial datum in H^3∩L^∞ there exists a smooth incompressible flow u such that the unique solution of (1.1) is global and belongs to C(R_+;H^3). The proof strategy is to first establish an L^∞-criterion (Proposition 3.1), then use the RAGE theorem for weakly mixing flows to obtain a local L^2 contraction of the deviation from the mean, iterate this contraction to make the L^2 deviation small, and finally use a nonlinear maximum principle on the torus to convert the resulting L^p control into a global L^∞ bound. The case β=d is treated in Section 3 and the case β∈[2,d) in Section 4.

Significance. If the argument were correct, the result would be a substantial improvement over previous work: it would extend mixing-induced suppression of blow-up for the generalized Keller-Segel system to the full range 0<α<2 and β∈[2,d], removing restrictions such as α>d/2 or the restriction to classical diffusion. The overall strategy, combining the RAGE theorem with the nonlinear maximum principle, is natural and attractive, and the appendix gives a self-contained proof of the nonlinear maximum principle on the torus. The main claim is clearly stated and falsifiable, and the paper does not rely on hidden fitting or ad-hoc numerical assumptions. However, the proof of the central global L^∞ estimate contains a load-bearing gap in the iteration of the mixing argument, so the theorem as stated is not established by the present manuscript.

major comments (3)
  1. [§3.2, Proposition 3.4, equations (3.39)–(3.51)] The iteration of the RAGE-based contraction is not justified. Lemma 3.5 is applied in (3.42) to the initial normalized fluctuation φ0=(ρ0−ρ̄)/‖ρ0−ρ̄‖_{L2}, and the membership φ0∈K is asserted immediately before (3.42). Even for j=0 this membership is not a consequence of the choice of N in (3.39); it is an additional condition that could be met by enlarging N, but the manuscript does not state this. More seriously, the 'repeat the above process k times' step after (3.51) requires φj=(ρA(jτ)−ρ̄)/‖ρA(jτ)−ρ̄‖_{L2}∈K for j=1,...,k−1. No argument for this membership is supplied. After the L2 contraction (3.49)–(3.51), the denominator ‖ρA(jτ)−ρ̄‖_{L2} shrinks, while the numerator is only known through an H^3 bound that itself depends on the L^∞ bound being proved; interpolation gives no uniform control of ‖φj‖_{\dot H^{α/2}} below λ_N^{α/2}. Since the contraction is iterated until ‖ρA(kτ)−ρ̄‖_{L2}≤B1, no fixed finite N can keep all iterates in the fixed compact set K. This gap is load-bearing: the global L^∞ estimate and hence Theorem 1.1 rest on it.
  2. [§3.2, equations (3.52)–(3.56) and final paragraph of Proposition 3.4] The passage from the single-time estimate (3.52) to the uniform L^p bound (3.53) is not justified as written. Inequality (3.52) gives ‖ρA(kτ)−ρ̄‖_{Lp}≤D only at the one time t=kτ, while the maximum-principle differential inequality (3.54) is integrated over the whole interval 0≤t≤kτ and therefore requires a bound on ‖ρA(t)‖_{Lp} for every t in that interval. The cited 'Theorem 2.6 and (3.52)' does not provide this: Theorem 2.6 is only a local well-posedness statement. On the first block one could obtain (3.53) from (3.41) via interpolation, but the text does not say this; for later blocks, where the same issue recurs in the sentence 'by the same argument with above', the missing uniform L^p or L^∞ control is exactly the quantity being proved. The final assertion that the same argument applies to the solution of (1.1) for all n∈Z+ is therefore unsupported.
  3. [§4.2, Proposition 4.4] Proposition 4.4 is proved by the one-sentence statement 'According to the proof of Proposition 3.4.' It therefore inherits verbatim the two gaps identified above. In addition, the equation (4.15) for the maximum contains the constant C0 multiplying the quadratic term, so the local time-scale estimates in Lemma 4.2 and the repeated contraction need to be rechecked with the modified constants; no such verification is given. Consequently, the proof of Theorem 1.1 for β∈[2,d) is not completed.
minor comments (5)
  1. [§3.2, (3.39)] The displayed condition for choosing N is notationally unclear: expressions such as '2400/23(C∞+ρ̄)' and '(1−B1^2/(B0^2−ρ̄^2))^{1/τ1}' should be written as explicit inequalities for λ_N^{α/2}, and the third condition should be λ_N^{α/2} ≥ (2/τ1)ln((B0^2−ρ̄^2)/B1^2).
  2. [§3.2, before (3.42)] The sentence 'Let (ρ0−ρ̄)/‖ρ0−ρ̄‖_{L2} ∈ K' should be stated as an additional condition on N and included in the selection of N; as written it appears as an assertion that is not implied by (3.39).
  3. [§4.2, Lemma 4.2, (4.17)] In the estimate of the term ρ̄∫(ρ−ρ̄)ΔK∗ρ dx, the displayed bound loses a factor of ‖ρ‖_{L∞}: one expects C0ρ̄‖ρ‖_{L∞}(‖ρ‖_{L∞}+ρ̄) rather than C0ρ̄(‖ρ‖_{L∞}+ρ̄). This affects the source term in (4.18).
  4. [§3.2, final paragraph of Proposition 3.4] The proof is carried out for the rescaled equation (3.37) with velocity Au, while Theorem 1.1 is stated for (1.1) with velocity u. Since the theorem allows choosing u, this can be repaired by declaring the final flow to be A times a fixed weakly mixing flow and noting that the weakly mixing property is preserved under time rescaling, but the passage is not explained.
  5. [Throughout] There are numerous typographical errors, including 'location solution' for 'local solution', 'we definite' for 'we define', 'downward rectification' for 'floor', and 'Combing' for 'Combining'. These should be corrected in revision.

Circularity Check

2 steps flagged · score 6.0 of 10

The proof of the global L∞ bound extends to all times by setting the input L∞ norm C∞ equal to the target CL∞, and by reapplying RAGE to later iterates whose compact-set membership is never established.

  1. self definitional [Section 3.2, Remark 9 (after the proof of Proposition 3.4)]
    "Remark 9. Without loss of general, we can assume C∞ = CL∞ for the completeness of proof."

    CL∞ is the global L∞ bound the proposition has to prove, while C∞ is the input bound ‖ρ0‖L∞ used in Lemma 3.2, in the choice of N in (3.39), and in the constants of (3.41)-(3.52). The final paragraph extends the estimate to all n∈Z+ by 'the same argument', which needs the local estimates at each block with an input L∞ bound. Replacing that input by CL∞ makes the input equal to the output, so the induction assumes the conclusion. The circularity is not cosmetic: CL∞ is only defined after the first block via (3.53)-(3.56), and the constants in (3.39) cannot be chosen before CL∞ is known.

  2. other [Section 3.2, proof of Proposition 3.4, after (3.51)]
    "and repeat the above process k times, we have"

    The process being repeated starts each substep by applying Lemma 3.5 to the normalized fluctuation (ρA(jτ)−ρ̄)/‖ρA(jτ)−ρ̄‖L2, which must belong to the compact set K = {φ∈S : ‖φ‖²_{H^{α/2}} ≤ λ_N^{α/2}}. The proof verifies this membership only for j=0 ('Let (ρ0−ρ̄)/‖ρ0−ρ̄‖L2 ∈ K'); the L2 contraction in (3.49)-(3.51) shrinks the denominator, so membership is not inherited. The only stated source of higher-order control for the iterates is Proposition 3.1, whose hypothesis is exactly the L∞ bound being proved, so the repeated RAGE application assumes the control that the argument is meant to establish.

full rationale

The paper does not fit parameters to data, rename a known result, or rely on the authors' own previous theorems: Lemma 2.1, Lemma 2.5, Lemma 3.5, Theorem 2.6 and Lemma 5.1 are all imported from independent external sources ([11,13,16,25,26,28] and [9,12,23]). The local H3 criterion and the nonlinear maximum principle are substantial and are not self-citations. The circularity is concentrated in the final bootstrap: Remark 9 explicitly identifies the target bound CL∞ with the input bound C∞, and the iteration 'repeat the above process k times' re-applies the RAGE lemma to later normalized iterates whose membership in the compact set K is never proved and, through (3.49)-(3.51), is only derivable from the L∞ bound under proof. Proposition 4.4 inherits both issues verbatim ('According to the proof of Proposition 3.4'). Thus the central global L∞ estimate, as written, depends on its own conclusion; this is a partial circularity rather than a mere technical omission.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The proof rests on standard functional-analytic machinery (positivity lemma, Sobolev inequalities), on imported dynamical results (RAGE theorem, weakly mixing flows), and on the nonlinear maximum principle whose appendix proof contains an unsupported WLOG step. There are no data-fitted constants and no invented entities; the hand-chosen parameters A, N, p, B1 are proof constants, not empirical fits.

free parameters (4)
  • A (flow amplitude) = chosen large: A ≥ A2(Tc, ρ0, τ1)
    Coupling constant for the mixing flow in (3.37); chosen large so the comparison estimate (3.46) and the substep condition τ = Tc/A ≤ τ1 hold.
  • N (spectral cutoff) = λ_N^{α/2} satisfying (3.39)
    Defines P_N and the compact set K via (3.40); all mixing averages and contraction factors depend on N.
  • p (Lp exponent) = p > d/α
    Free exponent; the nonlinear maximum principle ODE (3.55) needs 1 + pα/d > 2.
  • B1 (L2 deviation target) = min{(B0^2 − ρ̄^2)^{1/2}, (D/(2C∞ + ρ̄)^{1−2/p})^{p/2}}
    Stopping level for the block-wise contraction; used in (3.39) and (3.52) to guarantee the Lp budget D at block ends.
assumptions (7)
  • standard math Positivity lemma: (2/p)∫((-Δ)^{α/4}|f|^{p/2})² ≤ ∫|f|^{p-2}f(-Δ)^{α/2}f
    Cited as Lemma 2.1 from [13,26]; used for the dissipation terms in (3.21) and (3.30).
  • domain assumption Transport estimate: ‖ω(t)‖_{H^{α/2}} ≤ F(t)‖ρ0‖_{H^{α/2}} with D(t) ≤ C‖(-Δ)^{(2α+d+2)/4}u‖_{L2}
    Lemma 2.5, cited from [25,28]; requires u smooth and divergence-free; feeds the comparison estimate in Lemma 3.3.
  • domain assumption RAGE-type theorem (Lemma 3.5): for unitary U with purely continuous spectrum and compact K ⊂ S, (1/T)∫_0^T ‖P_N U^t φ‖² dt ≤ σ
    Cited from [11,16,28]; the engine converting weak mixing into time-averaged decay of low-frequency projections; applied with U^{At}.
  • domain assumption Existence of smooth weakly mixing incompressible flows on T^d
    Required for 'there exists u' in Theorem 1.1; cited to [20,21] (Fayad); weak mixing implies relaxation enhancing (Remark 3).
  • domain assumption Local well-posedness in H^3 with L1 conservation (Theorem 2.6)
    Proof is 'standard and similar to [1,29]', not given; the paper's global result builds on it.
  • domain assumption Nonlinear maximum principle on T^d (Lemma 5.1)
    Proved in the appendix following [9,12,23], but with an unjustified 'WLOG M ≥ 1/4' step about the distance from the maximum point to the boundary; the bound (3.17)-(3.18) is load-bearing for the L∞ estimate.
  • standard math Gagliardo-Nirenberg, Sobolev, Hölder, Poincaré inequalities on T^d
    Used throughout Sections 3-4 for the higher-order energy estimates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Suppression of blow up by mixing in generalized Keller-Segel system with fractional dissipation." pith.science (2026). https://pith.science/paper/QNKYTQ3H

@misc{pith2026190802489,
  author       = {Pith},
  title        = {Pith review of: Suppression of blow up by mixing in generalized Keller-Segel system with fractional dissipation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QNKYTQ3H}},
  note         = {Machine review of arXiv:1908.02489}
}
abstract

In this paper, we consider the Cauchy problem for a generalized parabolic-elliptic Keller-Segel equation with fractional dissipation and the additional mixing effect of advection by an incompressible flow. Under suitable mixing condition on the advection, we study well-posedness of solution with large initial data. We establish the global $L^\infty$ estimate of the solution through nonlinear maximum principle, and obtain the global classical solution.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 34 canonical work pages

  1. [1]

    Ascasibar, R

    Y. Ascasibar, R. Granero-Belinch´ on, and J. M. Moreno. An app roximate treatment of gravitational collapse. Phys. D , 262:71–82, 2013

    work page 2013

  2. [2]

    Bedrossian and S

    J. Bedrossian and S. He. Suppression of blow-up in Patlak-Keller- Segel via shear flows. SIAM J. Math. Anal. , 49(6):4722–4766, 2017

    work page 2017

  3. [3]

    B´ enyi and T

    A. B´ enyi and T. Oh. The Sobolev inequality on the torus revisited . Publ. Math. Debrecen, 83(3):359– 374, 2013

    work page 2013

  4. [4]

    Biler and G

    P. Biler and G. Karch. Blowup of solutions to generalized Keller-Seg el model. J. Evol. Equ. , 10(2):247– 262, 2010

    work page 2010

  5. [5]

    Biler and W

    P. Biler and W. A. Woyczy´ nski. Global and exploding solutions for n onlocal quadratic evolution problems. SIAM J. Appl. Math. , 59(3):845–869, 1999

    work page 1999

  6. [6]

    Blanchet, J

    A. Blanchet, J. Dolbeault, and B. Perthame. Two-dimensional Ke ller-Segel model: optimal critical mass and qualitative properties of the solutions. Electron. J. Differential Equations , pages No. 44, 32, 2006

    work page 2006

  7. [7]

    Bournaveas and V

    N. Bournaveas and V. Calvez. The one-dimensional Keller-Segel model with fractional diffusion of cells. Nonlinearity, 23(4):923–935, 2010. GENERALIZED KELLER-SEGEL SYSTEM 27

    work page 2010

  8. [8]

    M. P. Brenner, P. Constantin, L. P. Kadanoff, A. Schenkel, and S. C. Venkataramani. Diffusion, attraction and collapse. Nonlinearity, 12(4):1071–1098, 1999

    work page 1999

Show all 36 references
  1. [9]

    Burczak and R

    J. Burczak and R. Granero-Belinch´ on. Suppression of blow up b y a logistic source in 2D Keller-Segel system with fractional dissipation. J. Differential Equations , 263(9):6115–6142, 2017

    2017

  2. [10]

    A. P. Calder´ on and A. Zygmund. Singular integrals and periodic f unctions. Studia Math., 14:249–271 (1955), 1954

    1955

  3. [11]

    Constantin, A

    P. Constantin, A. Kiselev, L. Ryzhik, and A. Zlatoˇ s. Diffusion an d mixing in fluid flow. Ann. of Math. (2) , 168(2):643–674, 2008

    2008

  4. [12]

    Constantin and V

    P. Constantin and V. Vicol. Nonlinear maximum principles for dissipa tive linear nonlocal operators and applications. Geom. Funct. Anal. , 22(5):1289–1321, 2012

    2012

  5. [13]

    C´ ordoba and D

    A. C´ ordoba and D. C´ ordoba. A maximum principle applied to quas i-geostrophic equations. Comm. Math. Phys. , 249(3):511–528, 2004

    2004

  6. [14]

    Corrias, B

    L. Corrias, B. Perthame, and H. Zaag. Global solutions of some chemotaxis and angiogenesis systems in high space dimensions. Milan J. Math. , 72:1–28, 2004

    2004

  7. [15]

    Cusimano, F

    N. Cusimano, F. del Teso, L. Gerardo-Giorda, and G. Pagnini. D iscretizations of the spectral frac- tional Laplacian on general domains with Dirichlet, Neumann, and Rob in boundary conditions. SIAM J. Numer. Anal. , 56(3):1243–1272, 2018

    2018

  8. [16]

    H. L. Cycon, R. G. Froese, W. Kirsch, and B. Simon. Schr¨ odinger operators with application to quantum mechanics and global geometry . Texts and Monographs in Physics. Springer-Verlag, Berlin, study edition, 1987

    1987

  9. [17]

    Di Francesco, A

    M. Di Francesco, A. Lorz, and P. Markowich. Chemotaxis-fluid coupled model for swimming bacteria with nonlinear diffusion: global existence and asymptotic behavior. Discrete Contin. Dyn. Syst. , 28(4):1437–1453, 2010

    2010

  10. [18]

    R. Duan, A. Lorz, and P. Markowich. Global solutions to the cou pled chemotaxis-fluid equations. Comm. Partial Differential Equations , 35(9):1635–1673, 2010

    2010

  11. [19]

    L. C. Evans. Partial differential equations , volume 19 of Graduate Studies in Mathematics . American Mathematical Society, Providence, RI, second edition, 2010

    2010

  12. [20]

    B. Fayad. Smooth mixing flows with purely singular spectra. Duke Math. J. , 132(2):371–391, 2006

    2006

  13. [21]

    B. R. Fayad. Weak mixing for reparameterized linear flows on the torus. Ergodic Theory Dynam. Systems, 22(1):187–201, 2002

    2002

  14. [22]

    Friedman

    A. Friedman. Partial differential equations . Holt, Rinehart and Winston, Inc., New York-Montreal, Que.-London, 1969

    1969

  15. [23]

    Granero-Belinch´ on

    R. Granero-Belinch´ on. On a drift-diffusion system for semicon ductor devices. Ann. Henri Poincar´ e, 17(12):3473–3498, 2016

    2016

  16. [24]

    Hillen and A

    T. Hillen and A. Potapov. The one-dimensional chemotaxis model: global existence and asymptotic profile. Math. Methods Appl. Sci. , 27(15):1783–1801, 2004

    2004

  17. [25]

    Hopf and J

    K. Hopf and J. L. Rodrigo. Aggregation equations with fraction al diffusion: preventing concentration by mixing. Commun. Math. Sci. , 16(2):333–361, 2018

    2018

  18. [26]

    N. Ju. The maximum principle and the global attractor for the dis sipative 2D quasi-geostrophic equations. Comm. Math. Phys. , 255(1):161–181, 2005

    2005

  19. [27]

    Kiselev and L

    A. Kiselev and L. Ryzhik. Biomixing by chemotaxis and efficiency of b iological reactions: the critical reaction case. J. Math. Phys. , 53(11):115609, 9, 2012

    2012

  20. [28]

    Kiselev and X

    A. Kiselev and X. Xu. Suppression of chemotactic explosion by mix ing. Arch. Ration. Mech. Anal. , 222(2):1077–1112, 2016

    2016

  21. [29]

    D. Li, J. L. Rodrigo, and X. Zhang. Exploding solutions for a nonlo cal quadratic evolution problem. Rev. Mat. Iberoam. , 26(1):295–332, 2010

    2010

  22. [30]

    Li, J.-G

    L. Li, J.-G. Liu, and L. Wang. Cauchy problems for Keller-Segel t ype time-space fractional diffusion equation. J. Differential Equations , 265(3):1044–1096, 2018

    2018

  23. [31]

    Liu and A

    J.-G. Liu and A. Lorz. A coupled chemotaxis-fluid model: global ex istence. Ann. Inst. H. Poincar´ e Anal. Non Lin´ eaire, 28(5):643–652, 2011

    2011

  24. [32]

    A. Lorz. Coupled chemotaxis fluid model. Math. Models Methods Appl. Sci. , 20(6):987–1004, 2010

    2010

  25. [33]

    T. Nagai. Blow-up of radially symmetric solutions to a chemotaxis s ystem. Adv. Math. Sci. Appl. , 5(2):581–601, 1995. 28 BINBIN SHI AND WEIKE W ANG

    1995

  26. [34]

    Osaki and A

    K. Osaki and A. Yagi. Finite dimensional attractor for one-dime nsional Keller-Segel equations. Funk- cial. Ekvac. , 44(3):441–469, 2001

    2001

  27. [35]

    Senba and T

    T. Senba and T. Suzuki. Weak solutions to a parabolic-elliptic syst em of chemotaxis. J. Funct. Anal., 191(1):17–51, 2002

    2002

  28. [36]

    J. I. Tello and M. Winkler. A chemotaxis system with logistic source . Comm. Partial Differential Equations, 32(4-6):849–877, 2007. 1 School of Mathematical Sciences , Shanghai Jiao Tong Univer sity, Shanghai, 200240, P.R.China. E-mail address : binbinshi@sjtu.edu.cn 2 School o...

    2007

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.