Some tests methods fail randomly

[SergeStinckwich]

Some tests methods like :

• PMKolmogorovSmirnov1SampleTest >> testRejectOfEqualityHypothesesForSampleVersusDistribution
• PMBinomialGeneratorTest >> testBinomialGeneratorConvergesToMean
  fail randomly from time to time.

They should be rewritten in a better way.
Maybe we should control the seed in order to have reproducible test.

[hemalvarambhia]

This is a duplicate of #70, so I'll close that and this will be the new milestone issue.

[hemalvarambhia]

Can we do just Random seed: 0? To be honest this is a weak area for me in terms of testing.

[hemalvarambhia]

@SergeStinckwich In the case of the K-S test work in PR #111 one observation I made is that the one talks about about retaining or rejecting the null hypothesis, whereas the message rejectEqualityHypothesisWithAlpha doesn't use that language. This is also the case in wikipedia. So I was going to propose that we deprecate that in favour of retainNullHypothesisWithin: in a future version. What are your thoughts?