Idea: Preventing getting stuck with variable number of mutations

[perib]

TPOT2 can often get stuck without learning for long periods before finding a "breakthrough." One idea to potentially try to break out of this sooner could be to increase the number of mutations applied every generation in which there is no improvement in either the best score or the Pareto front.