Incremental evolution in ANNs: Neural nets which grow

This paper explains the optimisation of neural network topology using Incre mental Evolution; that is, by allowing the network to expand by adding to i ts structure. This method allows a network to grow from a simple to a compl ex structure until it is capable of fulfilling its intended function. The a pproach is somewhat analogous to the growth of an embryo or the evolution o f a fossil line through time, it is therefore sometimes referred to as an e mbryology or embryological algorithm. The paper begins with a general intro duction, comparing this method to other competing techniques such as The Ge netic Algorithm, other Evolutionary Algorithms and Simulated Annealing. A l iterature survey of previous work is included, followed by an extensive new framework for application of the technique. Finally, examples of applicati ons and a general discussion are presented.