Improve memory management through tree size estimation

[ktnr]

The required capacity is not known in advance, therefore we need an estimation of the (sub)tree size.

TwoStepBranchingProcedure/tsbp/src/BranchAndBound.cpp

Line 459 in d9850d1

////this->tree.m_vertices.reserve(250000000);

TwoStepBranchingProcedure/tsbp/src/BranchAndBound.cpp

Line 1822 in d9850d1

////this->tree.m_vertices.reserve(250000000);

Consider using the following relevant parameters:

• The epsilon parameter from Clautiaux, F., Carlier, J., & Moukrim, A. (2007). A new exact method for the two-dimensional orthogonal packing problem. European Journal of Operational Research, 183(3), 1196-1211, in combination with
• the distribution of items sizes, whereas the hardest instances are those where all items are smaller than half of the bin's width and height,
• item count, and
• the number of normal pattern placements points (which correlates with the homogeneity of items).

Generally, (significant) overestimation is fine, but might degrade performance for smaller instances.