Evolutionary Synthesis of Qca Circuits: a Critique of Evolutionary Search Methods Based on the Hamming Oracle

This paper introduces a discussion about evolutionary search methods based on Hamming oracle. In many optimization problems, the design of the fitness function includes the Hamming distance being referred this kind of functions as Hamming oracle. In this paper we adopt a critical look and ask ourselves to what extent genetic algorithms and other related evolutionary methods truly mimic evolution. We tested three evolutionary search methods taken as a case study the evolutionary synthesis of quantum-dot cellular automata circuits. Our main conclusion is that evolutionary search methods do not mimetic Darwinian evolution because knowledge is not obtained from the evolutionary surface exploration: evolution is the result of the ‘knowledge’ embedded by the researcher or human expert into the fitness function. Maybe a more appropriate denomination would be “combinatorial search algorithms" such as Minimax, Alpha-beta pruning, etc.