Refactorings for graded layer approach (2)

# Conflicts:
#	Core/Multilayer/Layer.cpp
#	Core/Multilayer/Layer.h
#	Core/Multilayer/MultiLayer.cpp
#	auto/Wrap/libBornAgainCore.py
#	auto/Wrap/libBornAgainCore_wrap.cpp

Making all parameters fittable.

Ensure that SimulationElements that are going to be added are always set (to zero if no calculation is done on a specific SimulationElement)

Removed some code duplication in MainComputation by handling ParticleLayoutComputation and RoughMultiLayerComputation uniformly

Refactor MainComputation: ParticleLayoutComputations are now in a simple vector (instead of vector<vector<..>>)