I have P number of vectors of n dimensions each, where n is real valued. I want to find the best vector out of P using the fitness function Z(P). I am not being able to think of the modelling of the mutation and crossover methods, which would result in valid offsprings (within P). How should I perform the crossover and mutation in this scenario ?