Abstract

One of the concerns in computer science involves optimizing usage of machines to make them more efficient and cost effective. One item of particular concern is the use of secondary storage devices, devices that store data other than in the main memory of the computer to which it is attached. The times for searching for data on these devices consistently proves to be a contributing factor in inefficient computer usage.

One data access method that avoids searching when possible is the hashing method. A function is defined to return the record number of a record based on its key field. The record can then be read in directly. A problem exists when more than one key maps to the same record number, called a collision, and must be dealt with, usually adding search time in the process.

Training a neural network to do this avoids these collisions. The Hamming network, based on the Hamming distances of two binary patterns, is trained to map the key fields directly to the record number of the data. The key must be converted to a binary format. The program passes the key to the network that simultaneously calculates the form of the Hamming distance between that key and all keys known to be in the file. A MAXNET network takes these distances and reduces them until no more than one is positive. The record number is found from the results, and the data can be accessed directly. All disadvantages from the software version are virtually eliminated.

Department(s)

Computer Science

Comments

This report is substantially the M.S. thesis of the first author, completed May 1990.

Report Number

CSc-90-4

Document Type

Article - Journal

Document Version

Final Version

File Type

text

Language(s)

English

Rights

© 1990 University of Missouri--Rolla, All rights reserved.

Publication Date

May 1990

Share

 
COinS