A gray box manifesto for evolutionary combinatorial optimization
Darrell Whitley · ACM SIGEVOlution · 2019
It is almost an article of faith that most evolutionary algorithms utilize random mutation and random recombination operators. Often the mutation is uniform random and sometimes the recombination operator is also a type of "Uniform Crossover." However when solving classic combinatorial optimization problems, random mutation and recombination operators are often both unnecessary and unproductive.