Global boundedness of a two-species attraction-attraction chemotaxis model with bilinear boundary influx

Technical Summary: Global Boundedness of a Two-Species Attraction-Attraction Chemotaxis Model with Bilinear Boundary Influx

Problem Statement and Motivation
The paper investigates the global existence and boundedness of classical solutions for a two-species chemotaxis system characterized by mutual attraction (attraction-attraction) and coupled with two chemical signals. While the classical Keller–Segel model typically employs homogeneous Neumann boundary conditions to ensure mass conservation, this study focuses on a scenario involving positive total flux across the boundary. Specifically, the model incorporates Robin-type boundary conditions where the inward flux depends on the product of cellular and chemical densities (bilinear influx).

The authors highlight that such boundary conditions, which generate inward fluxes proportional to $uv$ and $wz$, lead to a total mass that is not conserved but exhibits quadratic growth in time. This contrasts with previous studies on single-species models with subquadratic boundary influxes, where standard logistic damping sufficed for boundedness. The central mathematical challenge addressed is that, in this two-species setting with quadratic boundary influx, standard logistic damping ($au - bu^2$) is insufficient to counteract the mass increase. Consequently, the model requires gradient-dependent logistic damping terms (specifically $-c|\nabla u|^2$) to prevent finite-time blow-up.

Mathematical Model
The system under investigation consists of two cell populations, $u$ and $w$, and two chemical signals, $v$ and $z$. The equations are defined on a bounded smooth domain $\Omega \subset \mathbb{R}^n$:

1. Cellular Dynamics:
   $$u_t = \Delta u - \chi \nabla \cdot (u \nabla v) + a_1 u - b_1 u^2 - c_1 |\nabla u|^2$$
   $$w_t = \Delta w - \xi \nabla \cdot (w \nabla z) + a_2 w - b_2 w^2 - c_2 |\nabla w|^2$$
   The terms $-c_i |\nabla u|^2$ and $-c_i |\nabla w|^2$ represent the gradient-dependent damping mechanisms essential for the analysis.

2. Chemical Dynamics:
   The chemicals evolve according to either parabolic ($\tau=1$) or elliptic ($\tau=0$) equations:
   $$\tau v_t = \Delta v - v + w$$
   $$\tau z_t = \Delta z - z + u$$

3. Boundary Conditions:
   The system is subject to Robin-type conditions allowing inward flux:
   $$\partial_\nu u = (\alpha - 1)h_1 \chi u v, \quad \partial_\nu v = -h_1 v$$
   $$\partial_\nu w = (\alpha - 1)h_2 \xi w z, \quad \partial_\nu z = -h_2 z$$
   where $\alpha \in [0, 1]$. When $\alpha > 0$, the total flux is positive and quadratic in nature.

Methodology
The analysis proceeds through a rigorous sequence of a priori estimates and regularity arguments:

1. Local Existence: The authors first establish the local existence of classical solutions using fixed-point arguments in appropriate Hölder spaces, relying on standard parabolic and elliptic regularity theory.
2. Mass Control (L1 Estimates): A critical step involves deriving uniform-in-time bounds for the total mass $\int_\Omega (u + w)$. The authors demonstrate that under specific conditions on the damping coefficients ($b_i, c_i$) and the influx parameters ($\alpha, h_i, \chi, \xi$), the gradient-dependent damping terms effectively balance the quadratic growth induced by the boundary influx. Two distinct sets of sufficient conditions are derived:
   • One set requiring $b_i$ to be sufficiently large relative to $c_i$ and the boundary constants.
   • An alternative set involving a smallness condition on the product of the chemotactic sensitivity and the influx parameter ($\alpha \max\{\chi, \xi\}$).
3. Higher-Order Estimates (Lp Bounds): Building on the mass bounds, the authors derive $L^p$ estimates for $u$ and $w$ for arbitrary $p > 1$. This is achieved by testing the equations with powers of the solution variables and utilizing Gagliardo–Nirenberg interpolation inequalities. The presence of the $-c|\nabla u|^2$ term is crucial here, as it allows the absorption of the nonlinear boundary terms and the chemotactic cross-diffusion terms.
4. Extension to Global Boundedness: The paper employs a known extension criterion (Lemma 4.2), which states that if the $L^1$ norm of the solution and the $L^\infty$ norm of the gradient of the chemical signal are bounded, the solution remains globally bounded. The $L^p$ estimates derived in the previous step, combined with elliptic/parabolic regularity for the chemical equations, ensure that $\nabla v$ and $\nabla z$ remain bounded in $L^\infty$, thereby preventing blow-up.

Key Results
The paper establishes the following main theorems:

• Theorem 2.1 (Elliptic Case, $\tau=0$): Under the assumption that the domain is $C^{2+\delta}$, if the parameters satisfy either the "large damping" condition (inequalities involving $b_i, c_i, h_i, \chi, \xi$) or the "small sensitivity" condition, the system admits a unique global classical solution that is uniformly bounded in $L^\infty(\Omega \times (0, \infty))$.
• Theorem 2.2 (Fully Parabolic Case, $\tau=1$): Under the same parameter constraints (specifically the large damping condition), the fully parabolic system also admits a unique global bounded classical solution.

The results hold for both cases where the chemical equations are elliptic ($\tau=0$) and fully parabolic ($\tau=1$). The authors note that the mixed parabolic-elliptic cases ($\tau_1=0, \tau_2=1$ or vice versa) follow similar logic but present additional technical difficulties due to the coupling of different time scales, though the results are claimed to hold.

Significance and Claims
The authors position this work as a significant extension of recent single-species studies (specifically referencing [4]) to a two-species context. They emphasize that:

1. Non-Trivial Extension: The transition from a single-species to a two-species model is not straightforward. The interaction between the two populations and the mixed nonlinear terms involving both $u$ and $w$ introduce substantial analytical complexity that requires finer estimates than those used in single-species models.
2. Necessity of Gradient Damping: The study confirms that for two-species systems with quadratic boundary influx, standard logistic damping is insufficient. The inclusion of gradient-dependent damping terms ($-c|\nabla u|^2$) is mathematically necessary to restore control over the dynamics and ensure global boundedness.
3. Boundary Flux Complexity: The paper contributes to the emerging literature on chemotaxis models with non-conservative (inward flux) boundary conditions, distinguishing between subquadratic and quadratic influx behaviors and their respective requirements for damping mechanisms.

The paper concludes that while the framework generalizes previous single-species results, the two-species setting presents unique challenges that prevent a direct application of existing techniques, necessitating the development of specific a priori estimates tailored to the bilinear boundary influx and gradient-dependent damping structure.