A computational scheme based on random Boolean networks

Abstract

For decades, the size of silicon CMOS transistors has decreased steadily while their performance has improved. As the devices approach their physical limits, the need for alternative materials, structures and computational schemes becomes evident. This paper considers a computational scheme based on an abstract model of the gene regulatory network called Random Boolean Network (RBN). On one hand, our interest in RBNs is due to their attractive fault-tolerant features. The parameters of an RBN can be tuned so that it exhibits a robust behavior in which minimal changes in network's connections, values of state variables, or associated functions, typically cause no variation in the network's dynamics. On the other hand, a computational scheme based on RBNs seems appealing for emerging technologies in which it is difficult to control the growth direction or precise alignment, e.g. carbon nanotubes. © 2008 Springer Berlin Heidelberg.