GA: _calc_fitness_function() can return very large values which cause Overflow

[Gabriel-p]

I'm getting fitness values over 10000 for example, which cause the _select() function to overflow here:

        fitness[_idx] = math.exp(1 + fitness[_idx] * select_coeff)

Perhaps clip the fitness array before exponentiating?

fitness = np.clip(fitness, a_min=None, a_max=100)

[annoviko]

Hi, @Gabriel-p

Thank you for the reporting, the correction for the problem will be introduced.

[annoviko]

As I understand, the cause of the problem is non-normalized data. I would suggest to normalize data between [0, 1] to perform cluster analysis using genetic algorithm, for example, using the same approach as you suggested for the fitness function.

Nevertheless, I have added human-readable error that explains how to resolve the issue.