Abstract
dc:description.abstractThis thesis investigates a new approach to the optimization of functions using only data defined on sparse grids. The sparse grids provide a dramatic reduction in data compared to full grids, and spline interpolants are constructed to serve as accurate and smooth surrogates to the underlying functions being optimized. Using this approach, the number of grid points needed for approximation and optimization is significantly reduced without greatly compromising accuracy. In this thesis, we provide an overview of sparse grids, spline interpolation and optimization methods that are pertinent to our algorithm's development. We test several built-in optimizers in MATLAB, and implement our own conjugate gradient method using different choices for descent directions with an inexact line search based on the Armijo step length. We provide theoretical and numerical results using data from several test functions to demonstrate algorithm performance.
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Mathematics (M.S.)
- Level thesis:degree_level
- Thesis (restricted to Georgia Southern)
- Discipline thesis:degree_discipline
- Department of Mathematical Sciences
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ehirim, Chukwugozirim
- Contributors dc:contributor
-
- Emil Iacob
- Goran Lesaja
Subjects
dc:subject × 4Rights
- Licence dc:rights.license
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://digitalcommons.georgiasouthern.edu/etd/2253