Truncation error (numerical integration)

Truncation errors in numerical integration are of two kinds:

• local truncation errors – the error caused by one iteration, and
• global truncation errors – the cumulative error cause by many iterations.

Definitions

Suppose we have a continuous differential equation

$y'=f(t,y),\qquad y(t_{0})=y_{0},\qquad t\geq t_{0}$

and we wish to compute an approximation $y_{n}$ of the true solution $y(t_{n})$ at discrete time steps $t_{1},t_{2},\ldots ,t_{N}$. For simplicity, assume the time steps are equally spaced

$h=t_{n}-t_{n-1},\qquad n=1,2,\ldots ,N$

Suppose we compute the sequence $y_{n}$ with

$y_{n}=y_{n-1}+h\cdot A(t_{n-1},y_{n-1},h,f)$

Note: A is called an increment function, and can be interpreted as an estimate of the slope of y(t).

Local truncation error

The local truncation error is the error that our increment function, A, causes during a given iteration, assuming perfect knowledge of the true solution at the previous iteration.

More formally, the local truncation error, $\tau _{n}$, at step $n$ is defined by:

${\begin{aligned}\tau _{n}&=|y(t_{n})-y(t_{n-1})-h\cdot A(t_{n-1},y(t_{n-1}),h,f)|\end{aligned}}$

Global truncation error

The global truncation error is the accumulation of the local truncation error over all of the iterations, assuming perfect knowledge of the true solution at the initial time step.

More formally, the global truncation error, $e_{n}$, after $N$ steps is defined by:

${\begin{aligned}e_{N}&=|y(t_{N})-y_{N}|\\&=|y(t_{N})-(y_{0}+h\cdot A(t_{0},y_{0},h,f)+h\cdot A(t_{1},y_{1},h,f)+\cdots +h\cdot A(t_{N-1},y_{N-1},h,f))|\end{aligned}}$

Relationship between local and global truncation errors

Sometimes it is possible to calculate an upper bound on the global truncation error, if we already know the local truncation error. This requires our increment function be sufficiently well-behaved.

See also

• Numerical integration
• Numerical ordinary differential equations
• Truncation error

External links

• Notes on truncation errors and Runge-Kutta methods