Indexed text attachment
partial-differential-equations-in-higher-categories.tex
Source for Partial Differential Equations in Higher Categories
View paper · Download original file
Lines 2741–2840 of 3921. Source text is displayed literally.
2741\end{center}2742 2743\section{Laplace's equation}2744 2745The five-point scheme says that each value is the mean of its four neighbours: the discrete mean value property,2746with the random walk of \Cref{thm:kakutani} behind it \src{\S8.4}. The classical iterations solve it by running a2747process to equilibrium.2748 2749\begin{proposition}\label{prop:iter}2750\Scl[\cite{You50}] On an $N\times N$ grid the Jacobi iteration contracts the error by $\cos(\pi/N)$ per sweep, the2751Gauss--Seidel iteration by $\cos^2(\pi/N)$, and successive over-relaxation with the optimal parameter2752$\omega=2/(1+\sin(\pi/N))$ by $\omega-1$.2753\end{proposition}2754 2755\noindent For $N=16$ a random error is reduced by $10^{-6}$ in $585$, $298$ and $41$ sweeps. \chk{N6} Jacobi iteration is2756the explicit diffusion scheme at $s=\tfrac14$ run to its steady state, as the source notes; Gauss--Seidel is ``about2757twice as fast'', exactly so asymptotically; and the choice of $\omega$, which the source leaves as an art, has for2758this model problem the closed form above.2759 2760\section{Finite elements}2761 2762By \Cref{thm:dirichlet} the Galerkin solution in a space of trial functions is the energy projection of the true2763solution \src{\S8.5}. For piecewise linear functions on a uniform mesh of the interval the stiffness matrix is the2764difference Laplacian divided by $h$; the Galerkin solution is exact at the nodes, because the Green's function2765$x(1-y)$ is itself piecewise linear; and the energy error halves with $h$. \chk{N7} On a uniform triangulation of2766the square, linear elements give the five-point scheme. \chk{N8} The method earns its keep on domains where2767differences do not fit, and there the corner exponents of \Cref{prop:wedge} limit the rate.2768 2769\section{Discretize the complex}2770 2771\Cref{prop:tonti} suggests discretizing the two factors of the Laplacian separately: the exterior derivative2772exactly, on cochains of a mesh, and the Hodge star approximately.2773 2774\begin{theorem}[Finite element exterior calculus]\label{thm:feec}2775\Sli[\cite{AFW06}] On a triangulated domain the Whitney forms are a finite-dimensional subcomplex of the de Rham2776complex, isomorphic to simplicial cochains, and there are projections onto it that are bounded and commute with2777$\dd$. Consequently the discrete Hodge Laplacian is stable and convergent in every degree, and its harmonic forms2778have the dimensions of the cohomology of the domain.2779\end{theorem}2780 2781\noindent Stability here is a consequence of naturality: the discretization is a map of complexes. On a2782triangulated annulus with $12$ vertices, $24$ edges and $12$ triangles, $\dd\dd=0$ holds in integer arithmetic and2783the Betti numbers are $(1,1,0)$. \chk{N9} The single harmonic $1$-cochain is the discrete image of the class of2784$\dd\theta$, whose conjugate is $\log r$ (\Cref{prop:annulus}); a scheme that does not respect the complex can2785compute a spurious number of such modes.2786 2787\section*{Notes}2788 2789The source's Chapter 8 is followed for the schemes and their stability analysis. The equivalence theorem is2790\cite{LaRi56}; the stability condition for waves is \cite{CFL28}; the theory of over-relaxation is Young's2791\cite{You50}. Finite element exterior calculus is due to Arnold, Falk and Winther \cite{AFW06}.2792 2793\section*{Exercises}2794 2795\begin{exercise}\label{exer:amp}2796Derive the amplification factor of the explicit diffusion scheme and the condition $s\le\tfrac12$.2797\end{exercise}2798 2799\begin{exercise}\label{exer:sixth}2800Show that the explicit scheme has truncation error of fourth order in $\Delta x$ when $s=\tfrac16$.2801\end{exercise}2802 2803\begin{exercise}\label{exer:cflroots}2804For the wave scheme the amplification factors satisfy $\xi^2-2(1+p)\xi+1=0$ with $p=s(\cos k\Delta x-1)$. Show2805that both roots have modulus one for all $k$ exactly when $s\le1$.2806\end{exercise}2807 2808\begin{exercise}\label{exer:betti}2809Compute the Euler characteristic of the triangulated annulus of the text and deduce its first Betti number.2810\end{exercise}2811%% ---- end ch/12-numerics.tex ----2812\part{Physics and Nonlinearity}2813%% ---- begin ch/13-fields.tex ----2814\chapter{Fields}\label{ch:fields}2815 2816The source's chapter on physics \src{Ch.~13} is where the equations stop being scalar. Its topics are2817electromagnetism, fluids, scattering and the equations of particles. Each exhibits a structure that the scalar2818theory only hinted at, and one of them is the first object of this volume that is higher-categorical in earnest.2819 2820\section{Maxwell's equations}2821 2822In the source's units the equations are2823\[ \pd_tE=c\,\curl B-4\pi J,\qquad\pd_tB=-c\,\curl E,\qquad\Div E=4\pi\rho,\qquad\Div B=0. \]2824On space-time let $F=E\wedge dt+B$ be the field $2$-form. Up to constants and signs fixed by conventions, the2825second and fourth equations are $\dd F=0$ and the first and third are $\dd\star F=\star J$: two topological equations and one metric relation, the pattern of2826\Cref{prop:tonti}. Conservation of charge, $\pd_t\rho+\Div J=0$, is $\dd\dd=0$ applied to $\star F$. (The source's text2827gives this equation with the opposite sign of $\Div J$; its Exercise 1 of \S13.1 has it as here.) In vacuum each2828component satisfies the wave equation, and the transverse plane wave $E=(0,f(x-ct),0)$, $B=(0,0,f(x-ct))$ is a2829solution. \chk{P2}2830 2831Since $\dd F=0$, locally $F=\dd A$ for a potential $A$, determined up to $A\mapsto A+\dd\chi$.2832 2833\[\begin{tikzcd}[column sep=3.6em,row sep=2.4em]2834\underset{\text{gauge }\chi}{\Omega^0}\arrow[r,"\dd"] & \underset{\text{potential }A}{\Omega^1}\arrow[r,"\dd"] & \underset{\text{field }F}{\Omega^2}\arrow[r,"\dd"]\arrow[d,"\star"] & \Omega^3\\2835\Omega^4 & \underset{\text{current }\star J}{\Omega^3}\arrow[l,"\dd"'] & \Omega^2\arrow[l,"\dd"'] &2836\end{tikzcd}\]2837The top row is the de Rham complex of space-time and is natural for every smooth map; the bottom row is the same2838complex read from degree two; the one vertical arrow is the only place where the metric, and with it the speed of2839light, enters.2840