Theory of Natural and Artificial Evolution
Many people within (or closely associated with) the CCNR use Genetic Algorithms or Artificial Evolution as a methodology for designing complex systems of various kinds: by loose analogy with natural Darwinian evolution, using a population of trial solutions (represented by genotypes of `artificial DNA') to search through genotype space guided through successive generations by selection weighted towards increasing fitness. Two domains covered in different sections of this page are Evolutionary Robotics and Evolutionary Hardware. In parallel with such practical applications we have theoretical research on the dynamics of such population search.
One focus of research that brings together as many as 10 theoretical and practical researchers within the CCNR is Neutral Networks. It is suggested that a significant part of observed population dynamics is long periods when a population's fitness remains fairly static, even though at the genetic level there may be considerable change. Such a period of stasis may be terminated by a sudden increase in fitness -- `punctuated equilibrium'. The Neutral Network discussion group has a web page with a growing bibliography of papers relating to this area. For further questions please email Simon McGregor.