Hi, dear experts, Now I have some packing problems in E-commerce. Usually, some boxes (bins) of fixed size are given. Then the most suitable box (or a combination) is chosen to accommodate the items in order to minimize the material costs and logistic costs. This problem can be formulated as a standard 3D packing problem. Some heuristics are proposed, such as extreme points based method, the Guillotine algorithm, the maximal rectangles algorithm, etc. However, if we are given different size of boxes (bins), we may have different packing solutions and different costs. The costs (material costs and logistic costs, the latter is proportional to the number of boxed used) can be reduced by redesigning the box size. But how to find a better size is really complex. 3D-BP is already NP-hard. The method I can only put forward is to constantly call our packing algorithms, for different box-combinations. However, it is really time-consuming. Do you have some ideas to redesign the boxes to minimize the costs when packing given items? Any heuristics or mathmetical programming based methods are greatly appreciated. Thanks.
asked
LinYuan |

Sounds like an interesting topic. Can you contact me via email: operation.research100@gmail.com and we can discuss more details, i might be able to help you.

The purpose of this site is to provide some answers so all can benefit.

@opti100 Thanks. I am glad that you have the same insterest. Can you share your ideas in this site as @Mike Trick suggests? Thanks a lot. Looking forward to your reply.

I would try to build a mip, but it sounds to me more as a research project, than something that can be answered in a post. There are open questions like how are the material costs measured (volume/surface?), can the bins be made in strange sizes, which wouldn't be acceptable (i.e. very thin, but very high width/height). Then you need some statistics about what are the most common combination of items and what is the distribution etc.