Data Structures and Algorithms


The Backward Differentiation of the Bordering Algorithm for an Indefinite Cholesky Factorization

Authors: Stephen P. Smith

The bordering method of the Cholesky decomposition is backward differentiated to derive a method of calculating first derivatives. The result is backward differentiated again and an algorithm for calculating second derivatives results. Applying backward differentiation twice also generates an algorithm for conducting forward differentiation. The differentiation methods utilize three main modules: a generalization of forward substitution for calculating the forward derivatives; a generalization of backward substitution for calculating the backward derivatives; and an additional module involved with the calculation of second derivatives. Separating the methods into three modules lends itself to optimization where software can be developed for special cases that are suitable for sparse matrix manipulation, vector processing and/or blocking strategies that utilize matrix partitions. Surprisingly, the same derivative algorithms fashioned for the Cholesky decomposition of a positive definite matrix can be used again for matrices that are indefinite. The only differences are very minor adjustments involving an initialization step that leads into backward differentiation and a finalization step that follows forward differentiation.

Comments: 19 Pages.

Download: PDF

Submission history

[v1] 2017-07-17 17:13:10 (removed)
[v2] 2017-07-19 13:54:31 (removed)
[v3] 2017-07-20 12:52:57 (removed)
[v4] 2017-08-13 18:09:20

Unique-IP document downloads: 93 times is a pre-print repository rather than a journal. Articles hosted may not yet have been verified by peer-review and should be treated as preliminary. In particular, anything that appears to include financial or legal advice or proposed medical treatments should be treated with due caution. will not be responsible for any consequences of actions that result from any form of use of any documents on this website.

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus