Error in xtol termination: axis 1 is out of bounds for array of dimension 1 in numpy.AxisError

[fracerma]

Description

I'm using NSGAII algorithm for a custom Mixed Variable Problem, with something like x1 = Choice(), x2 = Choice(), x3 = Integer() as variable types.

I've set the termination as:

termination = DefaultMultiObjectiveTermination(
                            xtol=1e-8,
                            cvtol=1e-6,
                            ftol=1e-6,
                            period= 20,
                            n_max_gen = 150)

and the algorithm as:

algorithm = NSGA2(  pop_size=50,
                                    sampling=MixedVariableSampling(),
                                    mating=MixedVariableMating(eliminate_duplicates=MixedVariableDuplicateElimination()),
                                    eliminate_duplicates=MixedVariableDuplicateElimination(),
                                    )

Debugging the code, I have that inpymoo.termination.xtol.py , prev and current have the same dimensions:

prev = [{'x1': 'choice1_1', 'x2': 'choice2_1', 'x3' : 2}]
curr= [{'x1': 'choice1_1', 'x2': 'choice2_1',  'x3' : 2}]

Then e != prev have just one dimension, making np.mean([np.sum(e != prev, axis=1) to fail.

This seems to be fixed removing axis=1

Environment

• OS: Ubuntu 20.04
• Python Version: 3.7
• Package Versions:
  • numpy: 1.21.6
  • pymoo: 0.6.0.1

Error Traceback

  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/optimize.py", line 67, in minimize
    res = algorithm.run()
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/algorithm.py", line 141, in run
    self.next()
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/algorithm.py", line 162, in next
    self.advance(infills=infills)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/algorithm.py", line 232, in advance
    self._post_advance()
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/algorithm.py", line 309, in _post_advance
    self.termination.update(self)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/termination.py", line 28, in update
    progress = self._update(algorithm)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/default.py", line 24, in _update
    p = [criterion.update(algorithm) for criterion in self.criteria]
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/default.py", line 24, in <listcomp>
    p = [criterion.update(algorithm) for criterion in self.criteria]
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/termination.py", line 28, in update
    progress = self._update(algorithm)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/robust.py", line 32, in _update
    perc = self.termination.update(algorithm)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/core/termination.py", line 28, in update
    progress = self._update(algorithm)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/delta.py", line 42, in _update
    delta = self._delta(prev, current)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/xtol.py", line 21, in _delta
    return np.mean([np.sum(e != prev, axis=1).max() / len(e) for e in current])
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/pymoo/termination/xtol.py", line 21, in <listcomp>
    return np.mean([np.sum(e != prev, axis=1).max() / len(e) for e in current])
  File "<__array_function__ internals>", line 6, in sum
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/numpy/core/fromnumeric.py", line 2260, in sum
    initial=initial, where=where)
  File "/home/user/miniconda3/envs/fatai/lib/python3.7/site-packages/numpy/core/fromnumeric.py", line 86, in _wrapreduction
    return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
numpy.AxisError: axis 1 is out of bounds for array of dimension 1

[blankjul]

Honestly, I think the xtol only makes sense for variables that can be represented as float variables. Can you check out the commit above and see if this works for you?

[fracerma]

Make sense and now it works