Publication: Efficient signed backward substitution for piecewise affine functions via path problems in a directed acyclic graph
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About
Impress

Efficient signed backward substitution for piecewise affine functions via path problems in a directed acyclic graph

- Part of a collection -
 

Author(s)
T. Bosse , R. Seidler , H. M. Bücker

Published in
Proceedings of the 2021 SIAM Conference on Applied and Computational Discrete Algorithms (ACDA21), July 19--21, 2021

Editor(s)
M. Bender, J. Gilbert, B. Hendrickson, D. B. Sullivan

Year
2021

Publisher
SIAM

Abstract
We introduce an efficient signed backward substitution for a highly-structured system. More precisely, the problem is to find a vector u of dimension s that solves the system of piecewise affine equations u = c + L|u|, where L is a strictly lower left triangular s×s matrix, c denotes a given vector of dimension s, and the notation |\cdot| indicates the component-wise absolute value of a vector. The novel approach is based on a Neumann series reformulation and attempts to exploit a high degree of parallelism. We provide an analysis of its parallel run-time and show that it is suited for large, sparse systems whose triangular matrix has a small switching depth. The general idea behind this approach which is also used in the convergence proof is based on modelling the switching depth by a graph theoretic model. The key observation is that the computation of the switching depth corresponds to a single-pair shortest path problem in a directed acyclic graph. The proposed method is implemented and numerically evaluated using several examples whose problem structures are representative of various applications in scientific computing.

AD Theory and Techniques
Combinatorics, Computational Graph, Piecewise Linear

BibTeX
@INPROCEEDINGS{
         Bosse2021Esb,
       author = "T. Bosse and R. Seidler and H. M. B{\"u}cker",
       title = "Efficient signed backward substitution for piecewise affine functions via path
         problems in a directed acyclic graph",
       booktitle = "Proceedings of the 2021 SIAM Conference on Applied and Computational Discrete
         Algorithms (ACDA21), July~19--21, 2021",
       editor = "M. Bender and J. Gilbert and B. Hendrickson and D. B. Sullivan",
       pages = "171--181",
       address = "Philadelphia, PA, USA",
       publisher = "SIAM",
       doi = "10.1137/1.9781611976830.16",
       abstract = "We introduce an efficient signed backward substitution for a highly-structured
         system. More precisely, the problem is to find a vector $u$ of dimension~$s$ that solves the system
         of piecewise affine equations $u = c + L|u|$, where $L$ is a strictly lower left triangular $s
         \times s$ matrix, $c$ denotes a given vector of dimension~$s$, and the notation $|\cdot|$
         indicates the component-wise absolute value of a vector. The novel approach is based on a Neumann
         series reformulation and attempts to exploit a high degree of parallelism. We provide an analysis of
         its parallel run-time and show that it is suited for large, sparse systems whose triangular matrix
         has a small switching depth. The general idea behind this approach which is also used in the
         convergence proof is based on modelling the switching depth by a graph theoretic model. The key
         observation is that the computation of the switching depth corresponds to a single-pair shortest
         path problem in a directed acyclic graph. The proposed method is implemented and numerically
         evaluated using several examples whose problem structures are representative of various applications
         in scientific computing.",
       year = "2021",
       ad_theotech = "Combinatorics, Computational Graph, Piecewise Linear"
}


back
  

Contact:
autodiff.org
Username:
Password:
(lost password)