XNeuroGene: a system for evolving artificial neural networks
C. Jacob, J. Rehder, J. Siemandel, A. Friedmann · 2002
Artificial nervous systems (ANS) controlling action coordination tasks are in general very difficult to design. This is partly due to the great number of parameters that can be optimized: neural net connectivity, neuron functionality, weight settings etc. Another problem is posed by the need to automatically construct neuron clusters (modules) which can subsequently be used as building blocks. We focus on the description of a graphical user interface and management system (XNeuroGene) for NeuroGene which automatically assembles and optimizes artificial nervous systems based on evolutionary/genetic programming methods.>