作者: Michael O’Neill , Conor Ryan , Maarten Keijzer , Mike Cattolico
关键词: Evolutionary algorithm 、 Computer science 、 Automatic programming 、 Genetic algorithm 、 Parse tree 、 Genetic programming 、 Theoretical computer science 、 Fitness function 、 Genetic program 、 Crossover 、 Artificial intelligence 、 Grammatical evolution
摘要: Grammatical Evolution is an evolutionary automatic programming algorithm that can produce code in any language, requiring as inputs a BNF grammar definition describing the output and fitness function. The utility of crossover GP systems has been hotly debated for some time, this debate also arisen with respect to Evolution. This paper serves continue analysis operator by looking at result turning off crossover, exchanging randomly generated blocks headless chicken-like crossover. Results show essential on problem domains examined. mechanism one-point discussed, resulting discovery interesting properties could yield insight into operator's success.