Files in this item

FilesDescriptionFormat

application/pdf

application/pdf9411730.pdf (3MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Least Relative Change Quasi-Newton Updates for Chemical Engineering Process Optimization
Author(s):Nguyen, Thai Dong
Doctoral Committee Chair(s):Stadtherr, Mark A.
Department / Program:Chemical Engineering
Discipline:Chemical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Engineering, Chemical
Operations Research
Abstract:Advances in supercomputer architectures and mathematics have enhanced the performance of the equation-based approach in chemical engineering process simulation and design problems. However, the progress in the area of large-scale optimization is achieved at a slower pace thanks to difficulties in approximation of the sparse symmetric Hessian matrix by using traditional sparse least absolute change quasi-Newton updates.
In this work, four new sparse Hessian updates are derived by minimizing the relative change in the norm of the matrix. The new updates preserve symmetry and sparsity pattern of the Hessian. All four new updates are shown to be invariant to variable scaling, a property possessed only by the good dense least absolute change updates. The secant equation is satisfied in two of the least relative change updates which require the solution for an additional linear system with the same sparsity pattern of the true Hessian. Several important issues in implementing a computer code for large-scale optimization are also discussed. Preliminary numerical experiments demonstrate better performance of the new least relative change updates to previous sparse symmetric updates.
Issue Date:1993
Type:Text
Description:111 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1993.
URI:http://hdl.handle.net/2142/72147
Other Identifier(s):(UMI)AAI9411730
Date Available in IDEALS:2014-12-17
Date Deposited:1993


This item appears in the following Collection(s)

Item Statistics