2MMR10-Palletization

For all your palletization needs.

Run with python palletization.py [options] [dataset]

General options:

• sort:xxx: sorts the items by metric xxx in {none, rand, volume, side}. Other than none none and random random, this sorts in decreasing order. This is relevant for greedy and for level_03.
• no_dots: visualization will not display extreme points.
• no_vis: visualization will not be created.
• histo: only creates histograms of the data, does not run algorithms.
• validate: checks which properties are satisfied by an input solution.

Greedy options:

• greedy_01, greedy_02, greedy_03: specifies to use greedy algorithm. First has no look-ahead, second has one-step look-ahead, third has one-step look-ahead with weighted counting.
• onequeue: when the greedy algorithm fails to place a box, it will be tried after all other boxes have been tried (put at end of queue) (default).
• twoqueue: when the greedy algorithm fails to place a box, it will be tried after the first successful placement of another box (kept in separate queue).
• weight_const, flat, weight_top, weight_center, weight_center_flat, weight_center_top, flat_top, weight_center_flat_top: specifies the weight / utility functions used in greedy_03. Constant corresponds to greedy_02, flat values low height on boxes, weight top values low top height, weight center values boxes far from center. The others correspond to linear combinations of these (each component with weight 1).

Layers options (only implemented in Layers branch):

• level_01, level_02, level_03: specifies to use layer approach. First tries all side-lengths as layer height, second does the same but allows for fixing an item in the layer, third constructs layer for a fixed item chosen by sorting the items by some metric (default decreasing volume).
• total, density: whether the layers approach should optimize a layer for total volume in the layer, or for density of the layer (default total).
• support: the layers approach enforces strong stability.
• freeform: the layers approach does not enforce strong stability (default).
• fix_p: the layers approach fixes an item in the to-be-generated layer (for level_01 and level_02) or fixes the height of the layer to the height of the item (level_03).

Examples:
palletization.py greedy_01 sort:volume twoqueue dataset.csv
palletization.py greedy_03 sort:volume twoqueue weight_top dataset.csv
palletization.py level_03 sort:volume density support fix_p dataset.csv
palletization.py validate solution.csv