[MikeTheGreat]

Well, if there's one company on Earth that's both incentivized to find an algorithm to efficient pack stuff into their shipping bins and also well-financed enough to actually figure out a good linear or quadratic-time algorithm to do so, it's definitely Amazon.

And once they do so they'll have solved two big problems! :)

[darksaints]

As someone who worked on logistics optimization algorithms for Amazon, I’ll just say that the one thing Amazon did best was have clueless upper management continuously make poor strategic decisions that continuously nullified all of their improvements from optimization.

[neilv]

What was the root cause of that? (Company consciously prioritized other things than those logistics optimizations? Individual incentives lead to management behavior that was against company's intent? Bad hiring and retention practices for upper management or whomever was informing them?)

[SllX]

They might have the ability to do so. The motivation? Well let me put it this way: I tried Amazon’s grocery delivery service, and stopped using it because everything—everything—kept arriving in its own individual bag regardless of whether it made any sense, so it was just a bunch of bags I had to carry upstairs. That bags also had no handles.

So they were optimizing for something, but it definitely wasn’t packaging efficiency.

[fc417fc802]

Or the alternative that I occasionally encounter with non-grocery items - giant heavy item and small delicate item placed together in same box that is far too large for the both of them. A token piece of packing paper or lone plastic bladder tossed in, free to move about. The entire contents bouncing around.

Another amusing one was when they packed a somewhat delicate pantry food item in a paper envelope. It arrived thoroughly crushed, exactly as one would expect.

[smelendez]

Yes. I recently tried ordering a standard cardboard tube box of oats from Amazon, and it arrived crushed and leaking in its presumably nonsterile paper envelope. They gave me a refund and told me to throw it out.

I think this would have been much less likely to happen without the envelope, since whoever packed the truck would intuitively pack the tube vertically.

[dpark]

Optimizing for dollar cost. Human time costs more than the extra packaging.

Results would doubtless be different if they were optimizing for minimal environmental impact or produced waste.

[andrewljohnson]

Bin packing is theoretically NP Hard but practically solved all the time on real world datasets.

[datadrivenangel]

Yeah but NP Hard bin packing doesn't usually include situations where a flat screen TV squished on top of pallets comes sliding out of the truck when you go to unload...

Breakage results.

[andrewljohnson]

This sounds like a constraint I model in my bin packing optimization routine.

[kjellsbells]

I see what you did there. Touché.