Paper 1, Section II, 38E

Numerical Analysis
Part II, 2015

(a) The diffusion equation

ut=x(a(x)ux) in 0x1,t0\frac{\partial u}{\partial t}=\frac{\partial}{\partial x}\left(a(x) \frac{\partial u}{\partial x}\right) \quad \text { in } \quad 0 \leqslant x \leqslant 1, \quad t \geqslant 0

with the initial condition u(x,0)=ϕ(x)u(x, 0)=\phi(x) in 0x10 \leqslant x \leqslant 1 and zero boundary conditions at x=0x=0 and x=1x=1, is solved by the finite-difference method

umn+1=umn+μ[am12um1n(am12+am+12)umn+am+12um+1n]m=1,2,,M\begin{array}{r} u_{m}^{n+1}=u_{m}^{n}+\mu\left[a_{m-\frac{1}{2}} u_{m-1}^{n}-\left(a_{m-\frac{1}{2}}+a_{m+\frac{1}{2}}\right) u_{m}^{n}+a_{m+\frac{1}{2}} u_{m+1}^{n}\right] \\ m=1,2, \ldots, M \end{array}

where μ=k/h2,k=Δt,h=1/(M+1),umnu(mh,nk)\mu=k / h^{2}, k=\Delta t, h=1 /(M+1), u_{m}^{n} \approx u(m h, n k), and aα=a(αh)a_{\alpha}=a(\alpha h).

Assuming that the function aa and the exact solution are sufficiently smooth, prove that the exact solution satisfies the numerical scheme with error O(h3)O\left(h^{3}\right) for constant μ\mu.

(b) For the problem in part (a), assume that there exist 0<a<a+<0<a_{-}<a_{+}<\infty such that aa(x)a+a_{-} \leqslant a(x) \leqslant a_{+}for all 0x10 \leqslant x \leqslant 1. State (without proof) the Gershgorin theorem and prove that the method is stable for 0<μ1/(2a+)0<\mu \leqslant 1 /\left(2 a_{+}\right).