Possibilities and Constraints of Basic Computational Units in Developmental Systems
Konstantinos Antonakopoulos, Gunnar Tufte · 2009
Artificial systems often target organisms or systems with some kind of functionality. Taking inspiration from the cellular nature of its biological counterpart or at a more abstract level, we are in a position to investigate developmental mappings through a cellular or a non-cellular approach. This would lead to a discrete separation of the intermediate phenotype and their computational structures. In this paper, we investigate sparsely connected, simple computational elements that exist in specific architectures (i.e., Boolean network, Cellular automata, Artificial neural network and Cellular neural network), used in developmental settings. In addition, we present an analysis of the possibilities and constraints involved in the development of such computational elements in the targeted architectures focusing on the form, functionality and the inherent biological properties. 1