Bruno Garcia

> The problem is having two ways to represent an invalid state, especially when they can be at odds with each other (in that case, .has_value() && !->IsValid()). Fair enough,...

> That sounds interesting. Could you provide a performance benchmark (or other relevant metrics) between the current implementation and the one suggested in this issue? Yeah! I'm doing it and...

> It would also be interested in a comparison with repeated random draws. So when the EA does e.g. 10000 fitness evaluations, compare that with choosing the best individual from...

@martinus Cool! Going to test it and post the results here! Thank you!

@martinus thanks for your suggestion. Testing the mutation I proposed vs. creating the new individuals randomly, I realized that the results were very similar. So, I changed the approach. Considering...

@Xekyo Thanks, Murch! > I would imagine that a smaller mutation rate and a larger population would be helpful Yeah, I'm gonna test with a smaller one, most algorithms use...

@Xekyo Yeah, I know different approaches for crossover, but I think I will have to implement them to analyse what is the best one for this scenario. Something that makes...

@Xekyo I am gonna implement various solutions (different crossover and mutation) to check if the computational effort is satisfactory. The complexity of an EA depends on its implementation. Thanks for...

@aureleoules more or less, hope to do more simulations soon. Do you have any suggestions to improve the idea?

Closing for now as I haven't been working on it for a while