Random Neural Network Approach in Distributed Database Management Systems

Abstract

In this paper, Random Neural Network (RNN) approach has been applied to the distributed database

design of technology-corridor prototype project for Avcýlar Campus of Istanbul University in Turkey.

This project includes university, industry and government collaboration. Here, we need a distributed

environment for designing sub databases and fragmenting them on the sites. Therefore, different

techniques are considered for a database fragmentation. When techniques are described, eight different

properties are controlled for database process behaviors. Fragmentation techniques are ordered for

each property. These orders help us to make decision about which fragmentation technique is the best

for distributed database system. Here RNN approach and Radial basis functions networks are used for

generalization of selection of partitioning techniques. Training data of Radial basis function networks and

RNN are provided from the programs, which are executing under Oracle database. In this paper, firstly

we used Neural Networks approaches at distributed environments for automatic database fragmentation

selection operation and designed two non-linear algorithms. Then, Random Neural Network Methods

have been applied to the same problem and obtained satisfactory results.

Key Words: Database, database design, distributed database, database fragmentation, neural

networks, radial basis function networks, random neural network

Keywords

• -