Optimisation of network structure using genetic techniques

Arguments contending that it is necessary to use structured neural networks for the solution of certain problem types are presented. Structure is imposed on connectivity, activation functions, and other parameters of the network to simultaneously optimize generalization ability and compactness of the network. An analogy is made between the development of biological nervous systems from their genetic coding and the generation of artificial neural networks from a parametric description. Experiments which use genetic techniques to optimize network structure for a specific class of problem are described. Results which demonstrate the effectiveness of genetic optimization of network specifications in comparison with other optimization techniques are given. The parallel asynchronous implementation of genetic algorithms on a Sun-3 network is briefly described