> how is cooperative coevolution a crossover-dependent approach?
It's not. What I meant was that CCEAs and univariate EDAs (and most forms of ACO) make the same basic linkage assumption inherent in crossover, indeed to even more of an extreme: that you can assemble individuals from bits and pieces scattered about the space without any dependence issues.
It's not. What I meant was that CCEAs and univariate EDAs (and most forms of ACO) make the same basic linkage assumption inherent in crossover, indeed to even more of an extreme: that you can assemble individuals from bits and pieces scattered about the space without any dependence issues.