xants

Is it true that genetic algorithms have the benefit of being able to find the global optima more consistently due to the incorporation of randomness in subsequent generations? Whereas DL models often get stuck at local optima?