Add basic unit tests for problem::normalized.

[neduard]

Unit test for problem::normalized.
Test that applying normalize to a problem with bounds (-1, 1) doesn't have any effect.
Ensure that no matter what the bounds, normalized.objfun(x) = original.objfun(denormalize(x))

This PR is part of issue #7

[neduard]

A couple of questions:

1. should we include tests for normalized::compute_constraints_impl() ?
2. normalized::denormalize(const decision_vector& x) does NOT check that x is actually normalized: https://github.com/esa/pagmo/blob/master/src/problem/normalized.cpp#L77-83 is this the correct behaviour?

[darioizzo]

1. Yes, same as objfun_impl
2. good point, as the method is public a check and a pagmo_throw should be there. On the other hand, it is used by the private methods compute_constraints_impl and objfun_impl so this would trigger a lot of checks ... I doubt it would see any performance impact in any realistic case ....

[krzysztof]

Merge the lines here 63 and 64 (I know I didn't give you a clear guidelines on character limit ;) )

Method/Functions commenting style is either: https://github.com/esa/pagmo/blob/master/src/problem/normalized.cpp#L35-L41
or triple slash like https://github.com/esa/pagmo/blob/master/src/problem/normalized.cpp#L58
Latter is fine for unit testing.

[landscape-bot]

Code quality remained the same when pulling 445d66c on neduard:unit_test_normalize into 3ee3b0a on esa:master.