刚刚收到的一个邮件 这几天看下这个问题

10th December, 2013 
 
 
We are a company specialising in product traceability solutions, and are also involved in supply chain 
automation. We have a computational problem we need solved for a real-world application. If you can provide 
an efficient and commercially viable algorithm for achieving this then we would like to speak with you, and 
could offer contract work to assist our engineering team in implementing a real-world application.  
 
 
Problem Description 
 
Think of this as a mix of Tetris and The Tower of Hanoi. 
 
•     There is a finite set of rectangular cartons of varying height, width, length and weight. These are 
identified by barcode. 
•     The total quantity and type of each carton in a given run is known in advance, but not the order in 
which they will arrive. They arrive randomly in single order, one at a time. 
•     We need to stack these cartons on a pallet (or pallets, if there are too many in the run to fit on a single 
pallet within the height or stability limit) in a manner that is efficient, stable, and allows for the 
heaviest items to be stacked at the bottom and the lightest at the top i.e. both volume and weight 
need to be taken into account. 
•     A robotic arm picks the cartons and can place them either on the working pallet, or into a “holding 
pallet” area as it arranges the items. Once the working pallet is loaded it is automatically moved away 
and the holding pallet becomes the new working pallet and vice versa. 
•     A second working pallet area could be used, but this is not preferable. 
•     Stability has priority over efficiency of space used; the challenge is to maximise efficiency of space 
while maintaining stability and doing so in as short a time as possible.  
•     A completed pallet must be stacked in no more than 10 minutes. 
 
 
Contact: 
 
James Booth, CEO, Reynolds Group Ltd. 
Email: [email protected] 
Phone: 09 622 3500