ABSTRACT
There have been a variety of approaches to understanding such global phenomena. Cellular automata and artificial life use an empirical trial-and-error approach to finding local rules, but provide no framework for engineering local rules for a specific goal [10]. Genetic algorithms and evolutionary approaches generate local rules without any understanding of the individual element behavior, making them difficult to verify and characterize. Approaches within the applications community have therefore focused on centralized control-and
search-based approaches [2, 13]. My work is part of a larger vision called Amorphous Computing [1], whose goal is to identify engineering principles for designing robust self-organizing that achieve pre-established goals.
