Final elite solution is sometimes incorrect

[JFKakaJFK]

It can be the case that the values for the best solution are reported incorrectly (best solution reported has worse fitness compared to the best fitness reported). Here is a minimal reproduction example (should work as long as the gomea library and numpy are installed using e.g. pip install git+https://github.com/CWI-EvolutionaryIntelligence/GOMEA.git@dev#egg=gomea):

import gomea
import numpy as np

from ast import literal_eval

class CustomRosenbrockFunction(gomea.fitness.BBOFitnessFunctionRealValued):
    def objective_function( self, objective_index, variables ):
        f = 0
        for i in range(len(variables)-1):
            x = variables[i]
            y = variables[i+1]
            f += 100*(y-x*x)*(y-x*x) + (1.0-x)*(1.0-x)
        return f

def run_gomea():
    dim = 10
    frv = CustomRosenbrockFunction(dim,value_to_reach=1e-6)
    lm = gomea.linkage.Univariate()
    rvgom = gomea.RealValuedGOMEA(
        fitness=frv,
        linkage_model=lm,
        lower_init_range=-115,
        upper_init_range=-100,
        max_number_of_populations=1,
        base_population_size=100,
        max_number_of_evaluations=10_000
    )
    result = rvgom.run()
    result.printAllStatistics()


    stats = result.getFinalStatistics()
    best_solution = np.array(literal_eval(stats["best_solution"]))

    best_fitness = stats["best_obj_val"]
    actual_fitness = frv.objective_function(0, best_solution)
    print("Best solution: ", best_solution)
    print("Expected fitness: ", best_fitness)
    print("Actual fitness: ", actual_fitness)
    assert np.allclose(actual_fitness, best_fitness)

if __name__ == "__main__":
    for _ in range(100):
        run_gomea()

Truncated output:

...
Best solution:  [ 0.4489464   0.11328248  1.10134688  0.86169825  0.72128235  0.33731554
 -1.02180791 -1.44331219 -3.22583439 14.04847148]
Expected fitness:  4601.251681722282
Actual fitness:  5057.507173058592
Traceback (most recent call last):
   assert np.allclose(actual_fitness, best_fitness)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError