LEARNING ALGORITHM OF EVOLUTIONARY AMORPHOUS TRANSPARENT CELLULAR NEURAL NETWORK

摘要

PURPOSE: A learning algorithm of evolutionary amorphous transparent cellular neural network is provided which provides the cellular neural network having an amorphus structure and transparent data propagating property by choosing a cellular automata of a basic paradigm of an artificial life for analysing and modeling a complex adaptation system. CONSTITUTION: An amorphous connection structure by using a cellular automata is the characteristic of the learning algorithm of neural network. The amorphous connection structure is a structural evolutionary system.