A new genetic approach for neural network design

Neuro-genetic systems, a particular type of evolving systems, have become a very important topic of study in evolutionary design. They are biologically-inspired computational models that use evolutionary algorithms (EAs) in conjunction with neural networks (NNs) to solve problems. EAs are based on natural genetic evolution of individuals in a defined environment and they are useful for complex optimization problems with huge number of parameters and where the analytical solutions are difficult to obtain. This work present an approach to the joint optimization of neural network structure and weights, using backpropagation algorithm as a specialized decoder, and defining a simultaneous evolution of architecture and weights of neural networks.