English

Nonlinear Function Inversion using k-vector

Data Structures and Algorithms 2020-04-07 v1

Abstract

This work introduces a general numerical technique to invert one dimensional analytic or tabulated nonlinear functions in assigned ranges of interest. The proposed approach is based on an optimal version of the k-vector range searching, an ad-hoc modification devised for function inversion. The optimality consists of retrieving always the same number of data (1,2,1,2,\dots) for a specified searching range to initiate the root solver. This provides flexibility to adapt the technique to a variety of root solvers (e.g., bisection, Newton, etc.), using a specified number of starting points. The proposed method allows to build an inverse function toolbox for a set of specified nonlinear functions. In particular, the method is suitable when intensive inversions of the same function are required. The inversion is extremely fast (almost instantaneous), but it requires a one-time preprocessing effort.

Keywords

Cite

@article{arxiv.2004.02342,
  title  = {Nonlinear Function Inversion using k-vector},
  author = {David Arnas and Daniele Mortari},
  journal= {arXiv preprint arXiv:2004.02342},
  year   = {2020}
}

Comments

25 pages, 9 figures