Studying probabilistic faults in evolved non-uniform cellular automata

We study the effects of random faults on the behavior of one-dimensional, non-uniform cellular automata (CA), where the local update rule need not be identical for all grid sites. The CA systems examined were obtained via an approach known as cellular programming, which involves the evolution of non-uniform CAs to perform non-trivial computational tasks. Using the "system replicas" methodology, involving a comparison between a perfect, non-perturbed version of the CA and a faulty one, we find that our evolved systems exhibit graceful degradation in performance, able to tolerate a certain level of faults. We then "zoom" into the fault-tolerant zone, where "good" computational behavior is exhibited, introducing measures to fine-tune our understanding of the faulty CAs' operation. We study the error level as a function of time and space, as well as the recuperation time needed to recover from faults. Our investigation reveals an intricate interplay between temporal and spatial factors, with the presence of different rules in the grid giving rise to complex dynamics. Studies along this line may have applications to future computing systems that will contain thousands or even millions of computing elements, rendering crucial the issue of resilience.