Alternative results depends on the ordering of the Lines in it's data storage

[greenscientist]

When we calculate alternative, we go thought the lines in the order they are in their container. If we change the way they are ordered. (Like if we keep the same line data, but change their uuid/uid for example), we will get different results.
This is probably due because we reach the maxAlternatives quickly and don't explore all the options.
We cannot be sure we have the best alternative that way. (probably not possible with the CSA algorithm.)