A2.19 B2.19

Numerical Analysis
Part II, 2003

(i) Explain briefly what is meant by the convergence of a numerical method for ordinary differential equations.

(ii) Suppose the sufficiently-smooth function f:R×RdRdf: \mathbb{R} \times \mathbb{R}^{d} \rightarrow \mathbb{R}^{d} obeys the Lipschitz condition: there exists λ>0\lambda>0 such that

f(t,x)f(t,y)λxy,x,yRd,t0\|\mathbf{f}(t, \mathbf{x})-\mathbf{f}(t, \mathbf{y})\| \leqslant \lambda\|\mathbf{x}-\mathbf{y}\|, \quad \mathbf{x}, \mathbf{y} \in \mathbb{R}^{d}, t \geqslant 0

Prove from first principles, without using the Dahlquist equivalence theorem, that the trapezoidal rule

yn+1=yn+12h[f(tn,yn)+f(tn+1,yn+1)]\mathbf{y}_{n+1}=\mathbf{y}_{n}+\frac{1}{2} h\left[\mathbf{f}\left(t_{n}, \mathbf{y}_{n}\right)+\mathbf{f}\left(t_{n+1}, \mathbf{y}_{n+1}\right)\right]

for the solution of the ordinary differential equation

y=f(t,y),t0,y(0)=y0,\mathbf{y}^{\prime}=\mathbf{f}(t, \mathbf{y}), \quad t \geqslant 0, \quad \mathbf{y}(0)=\mathbf{y}_{0},

converges.