Positivity and long-term behaviour of a diffusion model with measure-valued nonlocal reaction term

Technical Summary: Positivity and Long-term Behaviour of a Diffusion Model with Measure-Valued Nonlocal Reaction Term

Problem Statement
The paper investigates the well-posedness, positivity, and long-term asymptotic behavior of solutions to a one-dimensional diffusion equation on the real line $\mathbb{R}$ featuring a singular, nonlocal reaction term. The model is given by:
$$\partial_t v = D\partial_x^2 v + e\Psi_t[v]\delta_0, \quad v(0, x) = v_0(x)$$
where $D > 0$ is the diffusion coefficient, $\delta_0$ is the Dirac measure at the origin, and the intensity $e\Psi_t[v]$ depends nonlocally on the solution $v$. This setup serves as a simplified representation of control systems in bioscience and engineering where a structure (e.g., a blood vessel, a cell membrane, or a permeable tube) exchanges compounds with its environment based on "measurements" taken at a distance (specifically at $x = \pm R$) but intervenes (sources or sinks) only at the origin.

A primary challenge addressed is the positivity of solutions. In physical applications (e.g., cell density, chemical concentration), solutions must remain non-negative. While positivity is guaranteed for positive initial data if the feedback is non-negative ($a \leq 0$), the paper focuses on the more complex case of negative feedback ($a > 0$), where the system attempts to regulate the concentration based on remote measurements. The authors note that classical maximum principles for local reaction-diffusion equations are inapplicable due to the nonlocality and the singularity of the Dirac measure.

Methodology
The authors employ a rigorous analytical approach centered on the Laplace transform, leveraging the essentially linear character of the uptake term after linearization or specific functional choices. The methodology proceeds as follows:

1. Dimensionless Formulation: The system is non-dimensionalized to reduce the number of parameters, focusing on the dimensionless feedback coefficient $\hat{a} = aR/D$.
2. Mild Solutions and Well-posedness: The existence and uniqueness of mild solutions in the Sobolev space $H^1(\mathbb{R})$ are established using results from Amann and Quitner regarding semilinear equations with measure-valued sources.
3. Laplace Transform Analysis: The diffusion equation is transformed into the $s$-domain. The nonlocal reaction term leads to algebraic relations involving transfer functions. Specifically, the sum of the solution at the measurement points, $u_+(t) = u(t, 1) + u(t, -1)$, satisfies a renewal equation.
4. Transfer Function and Delay Equations: The core of the positivity analysis involves studying the inverse Laplace transform of a specific transfer function $P_a(s, \beta)$. By utilizing a transformation (Lemma 3.6), the problem is reduced to analyzing a simpler meromorphic function $\tilde{P}_a(s, \beta)$, which corresponds to the solution of a linear delay differential equation (DDE) with constant delay:
   $$y'(t) = -a y(t-1), \quad t > 1, \quad y(t) = 1/2, \quad t \in [0, 1]$$
5. Positivity via DDE Theory: The authors utilize known results on the positivity of DDE solutions (specifically Györi & Ladas) and provide an independent proof using Laplace transform techniques to determine the range of the parameter $a$ for which the solution remains non-negative.
6. Pointwise Positivity: To ensure positivity on the subdomain $\mathbb{R} \setminus (-1, 1)$ (the "environment" outside the structure), the authors prove that pointwise positivity on the boundaries $x = \pm 1$ is sufficient. This is achieved by analyzing the symmetry and monotonicity of the initial condition $u_0$.

Key Contributions and Results

• Well-posedness: Theorem 2.1 establishes that for $u_0 \in H^1(\mathbb{R})$ and forcing $\Phi \in L^\infty_{loc}$, the system possesses a unique mild solution in $C^0(\mathbb{R}_0^+, H^1(\mathbb{R}))$.
• Positivity under Monotonicity (Theorem 2.2): The paper derives a sufficient condition for the solution to remain non-negative on the subdomain $\mathbb{R} \setminus (-R, R)$ for all time. Specifically, if the dimensionless feedback parameter satisfies $\hat{a} \in [0, e^{-1}]$ and the initial condition $u_0$ is non-negative, increasing on $(-\infty, 0)$, and decreasing on $(0, \infty)$, then the solution is non-negative on the exterior domain.
• Positivity under Symmetry (Proposition 2.4): A similar positivity result is obtained if the initial condition is symmetric around the origin, even without strict monotonicity, provided $\hat{a} \in (0, e^{-1}]$.
• Critical Threshold: The analysis identifies $e^{-1} \approx 0.3679$ as a critical threshold for the feedback coefficient. If $a > e^{-1}$, the associated DDE solution (and consequently the system solution) can become negative or oscillatory. The paper notes that while $a \leq e^{-1}$ is sufficient, the exact sharpness of this bound for all initial conditions remains an open question, though numerical simulations suggest the region of positivity is slightly larger than the theoretical guarantee but bounded by a curve related to the first derivative of the transfer function.
• Asymptotic Steady State (Theorem 2.3): Under the condition that the forcing term $\Phi(t)$ converges to a limit $\Phi_\infty$ and satisfies an $L^1$ regularity condition ($\Phi - \Phi_\infty \in L^1$), the solution converges pointwise to a constant steady state $\frac{\Phi_\infty}{2\hat{a}}$ on the subdomain $\mathbb{R} \setminus (-R, R)$.
• Numerical Verification: Section 6 provides numerical simulations using the Bromwich integral to compute the inverse Laplace transforms. These simulations confirm the theoretical predictions, showing stable positive solutions for $a \leq e^{-1}$ and the onset of oscillations or negativity for larger $a$.

Significance and Claims
The paper claims to provide a novel analytical framework for studying diffusion models with nonlocal, singular reaction terms, which are common in biological and engineering control systems but difficult to analyze with classical PDE techniques.

• Methodological Innovation: The primary significance lies in the use of Laplace transform arguments to convert the temporal evolution problem into an algebraic and complex-analytic one, specifically linking the positivity of the PDE solution to the positivity of a related delay differential equation.
• Physical Relevance: The results offer concrete parameter regimes ($\hat{a} \leq e^{-1}$) and initial condition constraints (monotonicity or symmetry) that guarantee physically meaningful (non-negative) solutions in systems where a "control" mechanism acts at a point based on remote sensing.
• Limitations and Scope: The authors are modest about the scope, noting that the results apply to the one-dimensional reduction of more complex two- or three-dimensional systems. They explicitly state that extending these Laplace transform techniques to higher dimensions is difficult and is left for future work. The paper focuses on establishing the approach to proving positivity and the existence of steady states rather than providing a comprehensive classification of all possible behaviors for arbitrary initial data.

In summary, the work successfully bridges the gap between abstract measure-valued reaction-diffusion equations and practical control problems in bioscience, offering rigorous conditions under which such systems remain stable and physically valid.