Empirical Study of the Influences of Genetic Parameters in the Training of a Neural Network
This paper presents the empirical results achieved in the computer system ROBOTS. This program simulates a virtual world, where agents, called robots, interact with an environment. Each robot is controlled by a neural network. The evolution of the robot behaviour (which is determined by the variation of the weights in the neural network) is done using a genetic algorithm. We describe the conceptual model used in ROBOTS. We also show how the genetic parameters and the environment itself influence the robot’s adaptation to the environment.
KeywordsNeural Network Genetic Algorithm Virtual World Genetic Parameter Neural Network Architecture
Unable to display preview. Download preview PDF.
- R.K. Belew, J. McInerney, and N.N. Schraudolph. Evolving networks: Using the genetic algorithm with connectionist learning. In Proceedings of the Second Artificial Life Conference, pages 511–547. Addison-Wesley, 1991.Google Scholar
- D.E. Goldberg. Genetic Algorithms in Search, Optimisation and Machine Learning. Addison-Wesley, Reading, MA, 1989.Google Scholar
- J. Hertz, A. Krogh, and R. Palmer. Introduction to the Theory of Neural Computation. Addison-Wesley, Reading, MA, 1991.Google Scholar
- V. Porto, D. Fogel, and L. Fogel. Alternative neural network training methods. IEEE Expert, pages 16–22, June 1995.Google Scholar