Title page for ETD etd-02132004-164149


Type of Document Dissertation
Author Dooley, Rion
Author's Email Address deardooley@hotmail.com
URN etd-02132004-164149
Title A Distributed Quadtree Dictionary Approach to Multi-Resolution Visualization of Scattered Neutron Data
Degree Doctor of Philosophy (Ph.D.)
Department Computer Science
Advisory Committee
Advisor Name Title
Aiichiro Nakano Committee Co-Chair
John Tyler Committee Co-Chair
Ahmad el Amawy Committee Member
Bijay Karki Committee Member
S. S. Iyengar Committee Member
Jorges Morales Dean's Representative
Keywords
  • approximation
  • compression
  • dictionary
  • quadtree
Date of Defense 2004-01-27
Availability unrestricted
Abstract
Grid computing is described as dependable, seamless, pervasive access to resources and services, whereas mobile computing allows the movement of people from place to place while staying connected to resources at each location. Mobile grid computing is a new computing paradigm, which joins these two technologies by enabling access to the collection of resources within a user's virtual organization while still maintaining the freedom of mobile computing through a service paradigm. A major problem in virtual organization is needs mismatch, in which one resources requests a service from another resources it is unable to fulfill, since virtual organizations are necessarily heterogeneous collections of resources.

In this dissertation we propose a solution to the needs mismatch problem in the case of high energy physics data. Specifically, we propose a Quadtree Dictionary (QTD) algorithm to provide lossless, multi-resolution compression of datasets and enable their visualization on devices of all capabilities.

As a prototype application, we extend the Integrated Spectral Analysis Workbench (ISAW) developed at the Intense Pulsed Neutron Source Division of the Argonne National Laboratory into a mobile Grid application, Mobile ISAW. In this dissertation we compare our QTD algorithm with several existing compression techniques on ISAW's Single-Crystal Diffractometer (SCD) datasets. We then extend our QTD algorithm to a distributed setting and examine its effectiveness on the next generation of SCD datasets. In both a serial and distributed setting, our QTD algorithm performs no worse than existing techniques such as the square wavelet transform in terms of energy conservation, while providing the worst-case savings of 8:1.

Files
  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  Dooley_dis.pdf 16.95 Mb 01:18:29 00:40:21 00:35:19 00:17:39 00:01:30

Browse All Available ETDs by ( Author | Department )

If you have questions or technical problems, please Contact LSU-ETD Support.