Improved Majority Identification by the Coarsened Majority Automaton

Abstract

The initial majority identification task is a fundamental test problem in cellular automaton research. To pass the test, an automaton must evolve to a uniform configuration consisting of the state that was in the majority for any initial configuration, employing only its internal, local dynamics. It is known that no two-state automaton can perform the majority task perfectly. Thus, it is a matter of continuing interest to identify and analyze new automata with improved majority identification capability. Here, we show that a coarsened version of one of the best majority identifiers can out-perform its parent automaton while simultaneously reducing the associated computational costs.