Hacker News Reader

Comment by robomartin | Hacker News Reader

Do any of these services have the ability to optimize packaging of goods?

Say you sell physical goods. A customer buys multiple items. They each have known shipping dimensions.

The question is: How do you consolidate the shipment into one or more boxes in order to optimize for minimum shipping costs?

Keep in mind that minimum shipping costs could also include the cost of shipping supplies as well as packaging and handling labor.

The most basic problem is fitting n packages of m[i] dimensions into the available stock of standard boxes and pick the arrangement that provides the least postage cost.

View on HN

You realize this is basically the same as the NP-Hard Knapsack problem?

Right, but that wasn't the question. Having done a lot of work with genetic algorithms, among other things, I have no doubt that I could sit down and write a reasonably useful solution. However, it would be absolutely brilliant if this was already part of an existing shipping API.

I think it would be very useful if these shipping API's accepted package specifications and helped optimize final packaging while in the process of providing a quote. If you want to talk about differentiating a solution, this, most definitely, would.

It may be infeasible at massive scale, but many systems do it right now with approximation algorithms, heuristics, and integer programming. Fedex, Walmart, Amazon, and most other industrial or logistics corporations have teams of people with a background in discrete optimization or operations research building and improving similar systems.

I'm not aware of any public APIs for this at present.

I'm in charge of IT for an ecommerce company, and I just wrote code for this last week.

We wanted to speed things up by packing product directly into our shipping boxes instead of boxing it up after the fact, and it was a lot of fun solving this problem. I specifically wrote very generic code, so there's no reason why it couldn't run off of any list of products/boxes.

If there was interest into this sort of thing, it probably wouldn't be terribly difficult to turn it into some sort of API over the next few months.

Reply on news.ycombinator.com