An improved implementation of sparsity detection of sparse derivative matrices

Simple item page
dc.contributor.authorJesmin, Tasnuba
dc.contributor.authorUniversity of Lethbridge. Faculty of Arts and Science
dc.contributor.supervisorHossain, Shahadat
dc.date.accessioned2019-01-09T22:06:20Z
dc.date.available2019-01-09T22:06:20Z
dc.date.issued2018
dc.degree.levelMastersen_US
dc.description.abstractOptimization is a crucial branch of research with application in numerous domain. Determination of sparsity is a vital stream of optimization research with potentials for improvement. Manual determination of sparsity structure of Jacobian matrix for a large problem is complicated and highly error-prone. The main motivation of this research is to propose an efficient algorithm which can effectively detect and represent sparsity of unknown Jacobian matrices. Automated sparsity detection algorithms find an optimal or near-optimal solution, which reduces time and space complexity for large scale data. Our proposed approach efficiently generates symmetric pattern utilizing band matrix and reduces the number of gradient evaluation. For efficient solution, we integrate our approach with existing pattern detection process. Greedy coloring algorithm is used for column portioning and multilevel algorithm with voting scheme is implemented for detection of sparsity pattern. Finally, parallel computation is used to reduce processing time of the overall approach.en_US
dc.description.sponsorshipS.G.S. Deans Scholarship- University of Lethbridge, Alberta Innovates Technology Futures Graduate Student Scholarship(AITF)en_US
dc.embargoNoen_US
dc.identifier.urihttps://hdl.handle.net/10133/5266
dc.language.isoen_USen_US
dc.proquestyesNoen_US
dc.publisherLethbridge, Alta. : Universtiy of Lethbridge, Department of Mathematics and Computer Scienceen_US
dc.publisher.departmentDepartment of Mathematics and Computer Scienceen_US
dc.publisher.facultyArts and Scienceen_US
dc.relation.ispartofseriesThesis (University of Lethbridge. Faculty of Arts and Science)en_US
dc.subjectJacobiansen_US
dc.subjectCombinatorial optimizationen_US
dc.subjectSparse matrices -- Data processingen_US
dc.subjectGraph coloringen_US
dc.subjectParallel programs (Computer programs)en_US
dc.subjectMatix devrivativesen_US
dc.subjectsparse data structureen_US
dc.subjectCPR algorithmen_US
dc.subjectsparse derivative matricesen_US
dc.subjectJacobian matrixen_US
dc.subjectmultilevel algorithmen_US
dc.subjectparallel implementationen_US
dc.titleAn improved implementation of sparsity detection of sparse derivative matricesen_US
dc.typeThesisen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Jesmin_Tasnuba_MSc_2018.pdf
Size:
3.18 MB
Format:
Adobe Portable Document Format
Description:
Final thesis pdf version
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
3.25 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Arts and Science, Faculty of
University of Lethbridge Theses