Engineering Amorphous Computing Systems

Abstract

How does one engineer robust collective behavior from the local interactions of immense numbers of unreliable parts? On the one hand, emerging technologies like MEMS are making it possible to assemble systems that incorporate myriad of information-processing units at almost no cost: smart materials, self-assembling structures, vast sensor networks. On the other hand, the plummeting cost of ad-hoc wireless communication is realizing the idea of pervasive computing: the creation of environments saturated with wireless computing devices collectively providing services anytime and everywhere. We discuss organizing principles and programming methodologies for controlling such amorphous systems, by combining robust algorithms inspired by nature with computer science techniques for controlling complexity.