Fix: Ignore order on task pairing checks

[stefantds]

In problem 17.01, the problem statement doesn't mention any particular order of the output. Currently the test only accepts the given order in the test file. My personal understanding of the problem statement is that the task assignment should

• also accept a different order of assignments: e.g. [(0, 1), (2, 3)] and [(2, 3), (0, 1)] should be equivalent
• task 1 and task 2 should be interchangeable: e.g. (0,1) and (1,0) should be equivalent

The proposed fix implements a way to achieve the above-mentioned fixes.

Context: while working on https://github.com/stefantds/go-epi-judge I got to look at some Java test code a bit closer. I've found some small potential improvements. I am opening a separate PR for each. Please have a look if there is something that can be improved.