An Encoding Scheme for Cooperative Coevolutionary Feedforward Neural Networks

摘要

The cooperative coevolution paradigm decomposes a large problem into a set of subcomponents and solves them independently in order to collectively solve the large problem. This work introduces a novel encoding scheme for building subcomponents based on functional properties of a neuron. The encoding scheme is used for training feedforward neural networks. The results show that the proposed encoding scheme achieves better performance when compared to its previous counterparts.