Dr. Veerman had an interesting presentation about his search for CA that actually performing common calculations like addition, subtraction, and multiplication. He also presented some techniques for optimized exhaustive searches over rule spaces. As far as I could tell, they dealt with pre-selection of rules with certain properties that could be determined before running the CAs. He concluded that CA are probably not the best systems to perform arithmetical calculations of the type he investigated and wondered about the best ways to search for the most efficient simple programs for a given calculation. I found this question very thought-provoking.
Comments