- GAs are useful on environments where no gradient exists nor any approximation of one
- They are also considered to be the last resort for optimization - GAs are highly inefficient optimizers due to their blind and random exploration of the space and dependence on luck and heuristics