Genetic algorithms for optimal logical database design

The focus of this paper is database design using automated database design tools or more general CASE tools. We present a genetic algorithm for the optimization of (internal) database structures, using a multi-criterion objective function. This function expresses connicting objectives , reeecting the well-known time/space trade-oo. This paper shows how the solution space of the algorithm can be set up in the form of tree structures (forests), and how these are encoded by a simple integer assignation. Genetic operators (database transformations) deened in terms of this encoding behave as if they manipulate tree structures. Some basic experimental results produced by a research prototype are presented.