College of Mathematics and Statistics, Chongqing University, Chongqing 401331, China.
Department of Applied Mathematics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong.
Math Biosci Eng. 2023 Jan;20(2):2011-2038. doi: 10.3934/mbe.2023093. Epub 2022 Nov 11.
We consider the following chemotaxis-growth system with an acceleration assumption, \begin{align*} \begin{cases} u_t= \Delta u -\nabla \cdot\left(u \bw \right)+\gamma\xkh{u-u^\alpha}, & x\in\Omega,\ t>0,\ v_t=\Delta v- v+u, & x\in\Omega,\ t>0,\ \bw_t= \Delta \bw -\bw +\chi\nabla v, & x\in\Omega,\ t>0, \end{cases} \end{align*} under the homogeneous Neumann boundary condition for $u,v$ and the homogeneous Dirichlet boundary condition for $\bw$ in a smooth bounded domain $\Omega\subset\R^{n}$ ($n\geq1$) with given parameters $\chi>0$, $\gamma\geq0$ and $\alpha>1$. It is proved that for reasonable initial data with either $n\leq3$, $\gamma\geq0$, $\alpha>1$ or $n\geq4,\ \gamma>0,\ \alpha>\frac12+\frac n4$, the system admits global bounded solutions, which significantly differs from the classical chemotaxis model that may have blow-up solutions in two and three dimensions. For given $\gamma$ and $\alpha$, the obtained global bounded solutions are shown to convergence exponentially to the spatially homogeneous steady state $(m,m,\mathbf 0$) in the large time limit for appropriately small $\chi$, where $m=\frac1{|\Omega|}\jfo u_0(x)$ if $\gamma=0$ and $m=1$ if $\gamma>0$. Outside the stable parameter regime, we conduct linear analysis to specify possible patterning regimes. In weakly nonlinear parameter regimes, with a standard perturbation expansion approach, we show that the above asymmetric model can generate pitchfork bifurcations which occur generically in symmetric systems. Moreover, our numerical simulations demonstrate that the model can generate rich aggregation patterns, including stationary, single merging aggregation, merging and emerging chaotic, and spatially inhomogeneous time-periodic. Some open questions for further research are discussed.
\begin{align*} \begin{cases} u_t= \Delta u -\nabla \cdot\left(u \bw \right)+\gamma\xkh{u-u^\alpha}, & x\in\Omega,\ t>0,\ v_t=\Delta v- v+u, & x\in\Omega,\ t>0,\ \bw_t= \Delta \bw -\bw +\chi\nabla v, & x\in\Omega,\ t>0, \end{cases} \end{align*} 其中$u,v$满足齐次 Neumann 边界条件,$\bw$满足齐次 Dirichlet 边界条件,在光滑有界区域$\Omega\subset\R^{n}$($n\geq1$)中,参数为$\chi>0$,$\gamma\geq0$和$\alpha>1$。证明对于合理的初始数据,要么$n\leq3$,$\gamma\geq0$,$\alpha>1$,要么$n\geq4$,$\gamma>0$,$\alpha>\frac12+\frac n4$,系统存在全局有界解,这与经典趋化模型有很大的不同,经典趋化模型在二维和三维空间中可能存在爆炸解。对于给定的$\gamma$和$\alpha$,当$\chi$足够小时,在大时间极限下,所得到的全局有界解会指数收敛到空间均匀的稳定态$(m,m,\mathbf 0$),其中如果$\gamma=0$,则$m=\frac1{|\Omega|}\jfo u_0(x)$,如果$\gamma>0$,则$m=1$。在稳定参数区域之外,我们进行线性分析以确定可能的模式形成区域。在弱非线性参数区域中,通过标准的微扰展开方法,我们证明上述不对称模型可以产生叉形分岔,这在对称系统中通常会发生。此外,我们的数值模拟表明,该模型可以产生丰富的聚集模式,包括静止、单个合并聚集、合并和出现混沌以及空间不均匀的时周期。讨论了进一步研究的一些开放性问题。