next up previous
Next: Bibliography Up: Numerical solutions of PDEs Previous: Methods for Hyperbolic Equations

Additional Components in Standard Equations

If problems have to be modeled, the standard parabolic, hyperbolic or elliptic equations don't describe the real world problems. Additional factors and components have to be involved into the equation. For example, as already seen in the chemotaxis equation (17):
$\displaystyle \frac{\partial n}{\partial t}$ $\textstyle =$ $\displaystyle D\nabla^2n-\nabla\cdot(n\nabla u),$  

a diffusion of cells can be forced towards a chemical signal by introducing an extra term.

In general, the other terms can be functions of the seeked function $u$:

$\displaystyle u_t=u_{xx}+g(u).$      

In the case $g(u)=\mbox{exp}(u)$ the approximation for $u_l^{m+1}$ (32) becomes
$\displaystyle \frac{u_l^{m+1}-u_l^m}{k}$ $\textstyle =$ $\displaystyle D\frac{u_{l-1}^m-2u_l^m+u_{l+1}^m}{h^2}+\mbox{exp}(u_l^m).$ (37)

If additional terms are functions of different variables, the model becomes a system of partial differential equations.
A model with three PDEs which was published by (1) is the following:

$\displaystyle \frac{\partial n}{\partial t}$ $\textstyle =$ $\displaystyle d_n\nabla^2n\; -\; \gamma\nabla\cdot(n\nabla f),$  
$\displaystyle \frac{\partial f}{\partial t}$ $\textstyle =$ $\displaystyle -\; \eta\; mf,$  
$\displaystyle \frac{\partial m}{\partial t}$ $\textstyle =$ $\displaystyle d_m\nabla^2m\; +\; \alpha n\;
-\; \beta m.$  

With the approximations for $u_t$, $u_x$ and $u_{xx}$:

$\displaystyle u_t\approx\frac{u_l^{m+1}-u_l^m}{k}\; ,\hspace{1cm}
u_x\approx\fr...
...l-1}^m}{2h}\; ,\hspace{1cm}
u_{xx}\approx\frac{u_{l-1}^m-2u_l^m+u_{l+1}^m}{h^2}$      

we obtain the following numerical approximations:

$\displaystyle \Bigg(\;
\frac{\partial n}{\partial t}$ $\textstyle =$ $\displaystyle d_n\frac{\partial^2 n}{\partial x^2}-
\gamma\Big[\frac{\partial n...
...\frac{\partial f}{\partial x}+
n\frac{\partial^2 f}{\partial x^2}\Big]\; \Bigg)$  
       

\begin{eqnarray*}
\frac{n_l^{m+1}-n_l^m}{k}&=&
d_n\frac{n_{l-1}^m-2n_l^m+n_{l+1}...
...1}^m-2m_l^m+m_{l+1}^m\big)+
k\alpha\; n_l^m-k\beta\; m_l^m+m_l^m
\end{eqnarray*}


next up previous
Next: Bibliography Up: Numerical solutions of PDEs Previous: Methods for Hyperbolic Equations
Heiko Enderling 2003-11-13