New results for triangular reaction cross diffusion system (Q2348400): Difference between revisions

Let \(\Omega\) be a smooth bounded domain of \(\mathbb{R}^N\), \(N\geq 1\), and consider the reaction-diffusion system \[ \begin{aligned} \partial_t u - \Delta \left( d_u u + u \phi(v) \right) = u \left( r_u - r_a u^a - r_b v^b \right) \quad &\text{in } (0,\infty)\times\Omega, \\ \partial_t v - d_v \Delta v = v \left( r_v - r_c v^c - r_d u^d \right) \quad &\text{in } (0,\infty)\times\Omega,\end{aligned} \] supplemented with homogeneous Neumann boundary conditions and nonnegative initial data. Here, \(d_u\), \(d_v\), \(r_u\), \(r_v\), \(r_a\), \(r_b\), \(r_c\), \(r_d\), \(a\), \(b\), \(c\), and \(d\) are positive numbers while \(\phi\) is a nonnegative \(C^1\)-smooth function. The above system features more general reaction terms than the well-known Shigesada-Kawasaki-Teramoto system (corresponding to \(a=b=c=d=1\)) but has a diffusion matrix which is simpler as it has an upper triangular structure with no cross-diffusion in the second equation. The existence of global weak solutions is proved when either \(d<a\) or \(a\leq d\) with \(a\leq 1\) and \(d\leq 2\), and is supplemented with regularity results and estimates. The existence proof is performed by taking the fast reaction limit \(\varepsilon\to 0\) of the singularly perturbed system \[ \begin{align*}{ \partial_t u_A - d_A \Delta u_A & = u_A \left( r_u - r_a (u_A+u_B)^a - r_b v^b \right) + \frac{1}{\varepsilon} \left( k(v) u_B - h(v) u_A \right) , \cr \partial_t u_B - (d_A+d_B) \Delta u_B & = u_B \left( r_u - r_a (u_A+u_B)^a - r_b v^b \right) - \frac{1}{\varepsilon} \left( k(v) u_B - h(v) u_A \right) , \cr \partial_t v - d_v \Delta v & = v \left( r_v - r_c v^c - r_d (u_A+u_B)^d \right) , }\end{align*} \] in \((0,\infty)\times\Omega\), where \(h\) and \(k\) are \(C^1\)-smooth functions satisfying \[ d_A+d_B \frac{h(r)}{(h+k)(r)} = d_u + \phi(r)\;, \quad h(r)\geq h_0, \, k(r)\geq h_0, \] for \(r\geq 0\) and some \(h_0>0\). Estimates which do not depend on \(\varepsilon\in (0,1)\) are derived for solutions \((u_A^\varepsilon,u_B^\varepsilon,v^\varepsilon)\) to the previous system by entropy and duality techniques and provide the convergence as \(\varepsilon\to 0\) of \((u_A^\varepsilon+u_B^\varepsilon,v^\varepsilon)\) towards a weak solution \((u,v)\) of the original system.