Learning and Evolution

  • Authors:

  • Stefano Nolfi;Dario Floreano

  • Affiliations:

  • Institute of Psychology, National Research Council, Viale Marx 15, Roma, Italy. nolfi@ip.rm.cnr.it;Laboratory of Microcomputing (LAMI), Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland. Dario.Floreano@epfl.ch

  • Venue:
  • Autonomous Robots
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

In the last few years several researchers have resorted toartificial evolution (e.g., genetic algorithms) and learningtechniques (e.g., neural networks) for studying the interactionbetween learning and evolution. These studies have been conducted fortwo different purposes: (a) looking at the performance advantages obtained by combining these two adaptive techniques; (b)understanding the role of the interaction between learning andevolution in biological organisms. In this paper we describe some ofthe most representative experiments conducted in this area and pointout their implications for both perspectives outlined above.Understanding the interaction between learning and evolution isprobably one of the best examples in which computational studies haveshed light on problems that are difficult to study with the researchtools employed by evolutionary biology and biology in general. Froman engineering point of view, the most relevant results are thoseshowing that adaptation in dynamic environments gains a significantadvantage by the combination of evolution and learning. These studiesalso show that the interaction between learning and evolution deeplyalters the evolutionary and the learning process themselves, offeringnew perspectives from a biological point of view. The study oflearning within an evolutionary perspective is still in its infancyand in the forthcoming years it will produce an enormous impact onour understanding of how learning and evolution operate.