作者: Peter T. Hraber , James P. Crutchfield , Melanie Mitchell
DOI:
关键词: Cellular automaton 、 Constructive 、 Function (mathematics) 、 Algorithm 、 Lambda 、 Genetic algorithm 、 Edge of chaos 、 Computation 、 Computer science 、 Interpretation (logic)
摘要: Author(s): Mitchell, Melanie; Hraber, Peter; Crutchfield, James P | Abstract: We present results from an experiment similar to one performed by Packard (1988), in which a genetic algorithm is used evolve cellular automata (CA) perform particular computational task. examined the frequency of evolved CA rules as function Langton's lambda parameter (Langton, 1990), and interpreted his giving evidence for following two hypotheses: (1) able complex computations are most likely be found near ``critical'' values, have been claimed correlate with phase transition between ordered chaotic behavioral regimes CA; (2) When computation, evolution will tend select values close critical values. Our produced very different results, we suggest that interpretation original not correct. also review discuss issues related lambda, dynamical-behavior classes, computation CA. The main constructive our study identifying emergence competition strategies analyzing central role symmetries evolutionary system. In particular, demonstrate how symmetry breaking can impede toward higher capability.