Large Artificial Nerve Net (LANNET)

Abstract

This report describes the implementation of a high speed self-organizing system based on the reinforcement principle. The self-organizing binary logical network is used as the primary component in the system. The learning system is a 1024 decision element netwolk with a general purpose program to enable the operator to simulate a large number of problems to study machine learning. The simulation of a maze runner problem and the results of some preliminary machine evaluation are presented. The system was developed for the Aeronautical Systems Division's Electronic Technology Laboratory to study complex biological problems.