[BUG]: Problem with negative fitness

[lucianolorenti]

Description of the bug

Hello!
Thank you for the library, it's been great working with it.

I am having issues when running ABC.

When the fitness function is negative ABC fails with the following message

File  mealpy\optimizer.py:636, in Optimizer.get_index_roulette_wheel_selection(self, list_fitness)
    633     final_fitness = np.max(list_fitness) - list_fitness
    635 prob = final_fitness / np.sum(final_fitness)
--> 636 return int(self.generator.choice(range(0, len(list_fitness)), p=prob))

File _generator.pyx:745, in numpy.random._generator.Generator.choice()

ValueError: probabilities are not non-negative

Steps To Reproduce

I know the problem does not make so much sense but 🤷

from mealpy import IntegerVar, ABC
import numpy as np

lower_bound = 0
upper_bound = 500
rows = 11
columns = 3
target_matrix = np.random.randint(lower_bound, upper_bound, size=(rows, columns))


def objective_func(solution):
    return np.sum(solution.flatten() - target_matrix.flatten())


problem_dict = {
    "obj_func": objective_func,
    "bounds": IntegerVar(
        lb=[0]*11*3,
        ub=[500]*11*3
    ),
    "minmax": "max",
}


abc_model = ABC.OriginalABC(epoch=1000, pop_size=100, n_limits=50)
g_best = abc_model.solve(problem_dict)

Additional Information

I think I know where the prolem is. I am creating an PR so you can take a look