I'm experimenting with different schemes for numerically solving PDEs in Mathematica. I use NestList to implement the Euler Method for 1-D PDE, the code is concise:

forwardEuler[t0_, y0_, tend_, h_, func_] := 
Module[{steps = Floor[(tend - t0)/h], ics = {t0, y0}},
NestList[{#[[1]] + h, #[[2]] + h *func /. {t -> #[[1]], y -> #[[2]]}} &,
ics, steps]]

and it works well.

ListLinePlot[{Table[Evaluate[{t,DSolveValue[{y'[t] == (y[t] + t^2 - 2)/(t + 1), y[0] == 2}, y[t],t]}], {t, 0, 6, 0.2}], 
forwardEuler[0, 2, 6, 0.2, (y + t^2 - 2)/(t + 1)]},PlotStyle -> {Blue, {Dashed, Orange}}, PlotLegends -> {"Exact Solution", "Forward Euler"}]

enter image description here

However, now I want to use Euler Method and central finite difference to solve PDE like the following.

$$ \left\{\begin{array}{l} \frac{\partial u}{\partial t}-\frac{\partial^{2} u}{\partial x^{2}}=x \mathrm{e}^{t}-6 x, \quad 0<x<1, \quad 0<t \leqslant 1, \\ u(x, 0)=x^{3}+x, \quad 0 \leqslant x \leqslant 1, \\ u(0, t)=0, \quad u(1, t)=1+\mathrm{e}^{t}, \quad 0<t \leqslant 1 . \end{array}\right. $$

Is it still recommended to write an Euler method using NestList (or any function from the Nest family)? It has a simple form in 1-D, but I'd like to see it in 2-D or even 3-D.

Additional: I also write codes of Implicit Euler, Trapezoidal, and Heun for 1D equations:

backwardEuler[t0_, y0_, tend_, h_, func_, tol_] := Module[{func1},
  func1[{\[FormalT]_, \[FormalY]_}] := {\[FormalT] + h, NestWhile[N[\[FormalY] + h*func /. {t -> \[FormalT] + h, y -> #}] &, \[FormalY] + h*func /. {t -> \[FormalT], y -> \[FormalY]}, (Abs[#1 - #2] > tol) &, 2, 10]};
  NestList[func1[#] &, {t0, y0}, Floor[(tend - t0)/h]]]


  • 1
    $\begingroup$ Euler method etc. are methods that can be extended to work on ODE system naturally, but your implementation can not. What you really need to think about is: how to code Euler method etc. so they work on ODE system? ( A single ODE is just a special case of ODE system. ) Related: mathematica.stackexchange.com/a/274162/1871 $\endgroup$
    – xzczd
    Jan 10 at 10:00


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.