The Skyline algorithm for packing 2D rectangles

I wonder how well Montecarlo works with a problem like this (for the online version), where you try all the potential places and run simulations adding new random boxes similar to the one added so far in order to see which option allows for better outcome. For sure it's slow, yet it should work well, which is interesting per-se since it would show once more how you can come up with slow but good solutions just simulating stuff. But likely, this is how this heuristics, like Skyline, were found: analyzing the properties of the brute-forced best picks.

They do this in subsets of aerodynamics for optimizing the tradeoff between simulation speed and fidelity in cases where there will be a lot of runs with the same geometry but different parameters