Implementing a parallel constrained l_ 1 approximation algorithm

DOI10.1016/0167-8191(93)90010-IMaRDI QIDQ1802663zbMATH OpenFDO

Authors W. A. Essah, T. B. Boffey

Publication date 28 February 1994

Published in Parallel Computing (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/0167-8191(93)90010-I

zbMATH Keywords

comparison of methodsconstrained \(\ell_ 1\) approximation algorithmnumerical resultsparallel computationsimplex method

Mathematics Subject Classification ID

Algorithms for approximation of functions (65D15) Numerical mathematical programming methods (65K05) Parallel numerical computation (65Y05) Linear programming (90C05)

Review text

The aim of this paper is to describe a parallel implementation of the constrained \(\ell_ 1\) approximation algorithm of \textit{I. Barrodale} and \textit{F. D. K. Roberts} [SIAM J. Numer. Anal. 15, 603-611 (1978; Zbl 0387.65027)]. The mathematical basis of the Barrodale-Roberts (BR) algorithm is presented in the introduction. The second section deals with the BR strategy via examples. The third part is dedicated to parallel implementation. The fourth part deals with numerical results. Eight problems are solved and compared both by Simplex02 and by the authors' algorithm. The new algorithm seems to be very much faster than the standard simplex method.

Recommendations

• An Improved Algorithm for Discrete l₁ Linear Approximation
• Computational experience with an algorithm for discrete \(L_ 1\) approximation
• A Finite Smoothing Algorithm for Linear $l_1 $ Estimation
• An \(L_{1}\) estimation algorithm with degeneracy and linear constraints.
• scientific article; zbMATH DE number 4050841

This page was built for publication: Implementing a parallel constrained \(l_ 1\) approximation algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1802663)