> When you're solving a problem, how do you know if the answer you've found is correct? In many domains, there is a single correct answer. A mathematical function may have a global maximum or other well-defined attributes. However, other problems, like how a cell behaves in a petri dish, do not have clear solutions. Enter evolution, which does not design towards a known solution but optimizes around constraints.

What about situations when there are no real optima and minima, but only intransitive relations? So "rock/paper/scissor"-like scenarios[0]. Could you use genetic algorithms in any way in those circumstances?

https://aeon.co/essays/attempts-to-choose-the-best-life-may-...

Not sure if @vanderZwan meant the same, but what about scenarios in which your goal is to pick the best member of an already existing population where each member has a set of variables like agility, strength, speed. Would it be possible to find a pareto-optimal solution with genetic algorithms, maybe even with priority-parameters?