Sashka Davis

Abstract

The purpose of this thesis is to study the behavior of the Hu- Tucker algorithm for building Optimal Alphabetic Binary Search Trees (OABST), to design an efficient implementation, and to evaluate the performance of the algorithm, and the implementation. The three phases of the algorithm are described and their time complexities evaluated. Two separate implementations for the most expensive phase, Combination, are presented achieving 0(n2) and O(nlogn) time and 0(n) space complexity. The break even point between them is experimentally established and the complexities of the implementations are compared against their theoretical time complexities. The electronic version of "The Complete Works of William Shakespeare" is compressed using the Hu- Tucker algorithm and other popular compression algorithms to compare the performance of the different techniques. The experiments justified the price that has to be paid to implement the Hu- Tucker algorithm. It is shown that an efficient implementation can process extremely large data sets relatively fast and can achieve optimality close to the Optimal Binary Tree, built using the Huffman algorithm, however the OABST can be used in both encoding and decoding processes, unlike the OBT where an additional mapping mechanism is needed for the decoding phase.

Library of Congress Subject Headings

Computer algorithms; Data structures (Computer science); Information retrieval; Trees (Graph theory)

Publication Date

1998

Document Type

Thesis

Department, Program, or Center

Computer Science (GCCIS)

Advisor

Radziszowski, S.

Advisor/Committee Member

Anderson, P.

Advisor/Committee Member

Kitchen, A.

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QA76.9.A43 D38 1998

Campus

RIT – Main Campus

Download

Share

COinS