Searching databases using parallel genetic algorithms on a transputer computing surface

Abstract Although many organisations have large databases containing useful information, few believe they are using these to maximum effect. When databases are very large, such as those provided by satellite observations or in highly computerised industries such as banking, searching them to discover the inherent information becomes a problem in itself. The volume of data involved is often simply too large to be tractable in real time using single processor technology. In this paper we describe the use of a parallel architecture, the Meiko computing surface, in conjunction with techniques from artificial intelligence in finding optimised information inherent in such databases, and give results which show the value of our approach.