SELF ORGANIZING NETWORKS.

The report summarizes the program in the development of pattern recognition automata logical networks with the ability to adjust their internal logical structure based on experience in a problem environment. The current research is concerned with collections of individually fixed automata, which are regarded as a species. The adaptation or change takes place not within particular individual automata, but in terms of an evolutionary process which carries the species into harmony with its environment. The general observations which led to the present approach are outlined. The evolutionary model is described in detail. Some of the possible applications of an evolutionary search and some of the preliminary results are outlined. Author