Thank you for the response. Though this is interesting, I'm not worried about the ability to pack items efficiently, rather reducing the man power required to pick items. So if 10 of the 15 items required are already contained in a "master carton" then we pick 6 total boxes (one master carton and five individual boxes) rather than picking 15 total boxes.

So the algorithm would need to determine if the contents of a master carton will be 100% utilized to fulfill part of the order, then determine if there are other master cartons that could fill any of the remainder of the order, and continue on until no more master cartons will work, and return the remaining items to be picked.