Possible bug in PoolLast's is_empty method
In class PoolLast from shadeils.py we can find (line 73):
def is_empty(self):
counts = self.improvements.values()
return np.all(counts == 0)
Given the fact that the result of dictionary class's method values() returns some type of a collection,
no matter the contents of the dictionary, the comparison values() == 0 will always yield False.
As a result, if not pool_global.is_empty(): statement found in line 292 will always be executed .
I assume that the behaviour is unintended, but does it have an impact on the performance, overall logic of the method?
Thank you for detecting that. Fortunately, I think it is not relevant, because, as I remember, the experiments were carried out with a pool non empty. Anyway, I will check, and give you a more complete information (and a fix) later.
Thank you for your response!
I need SHADE-ILS for my own research, and because of that I have downloaded and started to organise your code.
I have noticed that a lot of it is not being used (got rid of ~3/4 of it at the moment), and removal of it had no impact on the method.
Would you mind if I shared refurbished code with you so you could update your own repository?
I need SHADE-ILS for my own research, and because of that I have downloaded and started to organise your code.
Ok.
I have noticed that a lot of it is not being used (got rid of ~3/4 of it at the moment), and removal of it had no impact on the method.
Yes, sometimes it happens. I write several options that finally did not work, and some code is actually not used at all. However, when I publish the paper, I prefer to submit the source code used for the experimental section, so sometimes that code was not removed.
Would you mind if I shared refurbished code with you so you could update your own repository?
I do not mind, however, in order to maintain localled the original source code, I prefer to have for now in another branch. Anyway, if you need any help (I have worked a lot in LSGO, give tutorials, and several colllaborations), does not hesitate in contact with me.
Thank you for your response!
I am waiting for your insights on the bug :)