Convergence of Iterative Methods for Quadratic Programing Problems

by Kristin Sickau
This project regards the study of the global convergence of iterative methods for quadratic programming. There are several iterative methods commonly used in quadratic programming, and our research focuses on Dykstra's cyclic projections method. We implemented a novel version of the Dykstra's algorithm in Python, which uses sparse matrices and is memory efficient. In addition, we include the variational inequalities problem as an application of Dykstra's algorithm.
Major: 
Mathematics
Exhibition Category: 
Physical Sciences
Exhibition Format: 
Poster Presentation
Campus: 
University Park
Faculty Sponsor: 
Ludmil Zikatanov
Poster Number: 
96