Computational Issues in High Performance Software for Nonlinear Optimization
- PDF / 161,543 Bytes
- 6 Pages / 595 x 842 pts (A4) Page_size
- 53 Downloads / 167 Views
#1998 Operational Research Society Ltd. All rights reserved. 0160-5682/98 $12.00 http://www.stockton-press.co.uk/jor
Book Selection Edited by JM Wilson A Murli and G Toraldo (Eds): Computational Issues in High Performance Software for Nonlinear Optimization JS Edwards and PN Finlay: Decision Making with Computers: The Spreadsheet and Beyond GB Dantzig and MN Thapa: Linear ProgrammingÐ1: Introduction NK Jaiswal: Military Operations Research: Quantitative Decision Making P van Hentenryck, L Michel and Y Deville: NumericaÐA Modeling Language for Global Optimization LH Arps, B Alderman, E Hewitt and GB Price (HJ Miser ed.): Operations Analysis in the Eighth Air Force, 1942±1945ÐFour Contemporary Accounts
Computational Issues in High Performance Software for Nonlinear Optimization A Murli and G Toraldo (Eds) Kluwer Academic Publishers, London, 1997. 159 pp. £81.75 ISBN 0 7923 9862 9 This book contains a selection of the papers presented at the conference `High Performance Software for Nonlinear Optimisation: Status and Perspectives', held in Capri (Italy) in 1995. It also appeared as a special issue of Computational Optimization and Applications. Thus this is more than a conference proceedings: the editors succeed in doing a good job of assembling, editing and refereeing the selected papers, so that the result is more homogeneous than what can be expected from general proceedings. The style of the papers is quite oriented towards computer implementation, with generally more details on low level algorithmic choices than usually found in mathematical programming journals. Most of the algorithms are introduced with plenty of details and quite extensive numerical tests. The book starts with a comparison of different algorithms for mixed complementarity by Billups, Dirkse and Ferris, where the authors compare several methods through a common GAMS interface and report quite extensive computational results. The second paper, by Bouaricha and More', describes an environment, ELSO, for the unconstrained minimization of large scale partially separable objective functions. The next paper, by Conn, Gould and Toint, gives a theoretical analysis of a method for large scale non linear optimisation based upon Langrangean relaxation. This paper contains no computational results, but its contents are connected with the development of the well known LANCELOT package Another paper, by Lucidi and Roma, reports extensive numerical tests on a class of truncated Newton methods for unconstrained optimization based upon curvilinear search paths. Next, a paper by Matstoms surveys available algorithms and software for
1225 1225 1226 1227 1228 1229
sparse linear least squares. A paper by Mavridou and Pardalos presents different heuristics for the facility layout problemÐthis is the unique paper in the volume which does not deal with classical non-linear optimisation: even if it is clear that combinatorial optimisation can be seen as a special case of nonlinear programming, this paper (which is indeed interesting in its own right) seems to be qui
Data Loading...
