ONLINE PACKING WITH BOXES OF DIFFERENT SIZES

The packing of boxes distributed to different customers from a central depot is under consideration in this study. The boxes are of various sizes and they arrive on a conveyor belt one at a time. The on-line si tuation is assumed, which means that at each stage we only know the la yout of the previous boxes on the partially filled pallet and the size of the box to be placed next but we have no information about the for thcoming ones. Further it is not allowed to move boxes already placed. The objective is to produce efficient and stable loads by an on-line packing algorithm, which can be implemented in automated systems. The algorithms we develop are based on the well-known first-fit and best-f it principles and their efficiency has been tested empirically.