Sequence transformation

from Wikipedia, the free encyclopedia

In mathematics, a sequence transformation is a transformation that is used to numerically calculate the limit value of a slowly convergent sequence or series , or the antilimes of a divergent series .

For a given sequence

is the transformed sequence

.

The elements of the transformed sequence are normally calculated as a function of a finite number of elements of the original sequence. So there is an illustration of the form

with a finite . In the simplest case, these and the are real or complex numbers . In general, these are elements of a vector space or an algebra .

The transformed sequence is said to converge faster than the original sequence if

where the (anti-) limes of is. If the original sequence converges slowly, this is called convergence acceleration .

If the mapping is linear in each argument, i. i.e., if

for constants

holds, then the sequence transformation is called a linear sequence transformation , otherwise a non-linear sequence transformation .

A sequence transformation can be used to accelerate the convergence of a convergent series or as a summation method for a divergent series : For a series

just look at the result

the partial sums

and applies a suitable sequence transformation to this.

Important examples of nonlinear sequence transformations are Padé approximants for power series and Levin-like sequence transformations .

Particularly non-linear sequence transformations often result in highly efficient extrapolation methods .

literature

  • C. Brezinski and M. Redivo Zaglia: Extrapolation Methods. Theory and Practice. North Holland, 1991.
  • GA Baker, Jr. and P. Graves-Morris: Padé Approximants. Cambridge UP 1996.