A new framework for global optimization : application to control of flexible structures /

Bibliographic Details
Main Author: Browder, Andrew Maxwell, 1961-
Other Authors: Alexander, Richard M. (degree committee member.), Kurdila, Andrew J. (degree committee member.), Vadali, Srinivas R. (degree committee member.)
Format: Thesis Book
Language:English
Published: 1993.
Subjects:
Online Access:ProQuest, Abstract
Link to OAKTrust copy
Description
Abstract:A new approach to the solution of global optimization problems is developed, and compared to several pre-existing optimization methods. The new global optimization approach embodies a framework of simultaneous, coordinated, global and local searches. This new approach is shown to have many advantages over competing methods: rigorously characterizable convergence properties; ability to efficiently locate multiple feasible designs; customizable to specific problems, i.e. to exploit known problem characteristics in obtaining improved convergence properties; effective utilization of high performance computing environments. Analytical models of the global optimization process are established to enable a unifying examination of both deterministic and stochastic approaches to global optimization. Two currently popular stochastic optimization methods, namely, genetic (or evolution) algorithms, and simulated annealing, are shown to be subclasses of more general (and more practical) optimization approaches. A family of test functions is created and used in numerical studies, comparing two instances of the new approach to both simulated annealing and genetic algorithms. To demonstrate the application of the new approach to a realistic optimization problem, sensor/actuator placement optimizations, as well as control gain design optimizations, are conducted for a representative flexible structure.
Item Description:Vita.
"Major subject: Aerospace Engineering."
Physical Description:xi, 189 leaves : illustrations ; 28 cm
Bibliography:Includes bibliographical references.