Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10347569 | Computers & Operations Research | 2013 | 12 Pages |
Abstract
The problem of packing two-dimensional items into two-dimensional bins is considered in which patterns of items allocated to bins must be guillotine-cuttable and item rotation might be allowed (2BP|â|G). Three new constructive heuristics, namely, first-fit insertion heuristic, best-fit insertion heuristic, and critical-fit insertion heuristic, and a new justification improvement heuristic are proposed. All new heuristics use tree structures to represent guillotine-cuttable patterns of items and proceed by inserting one item at a time in a partial solution. Central to all heuristics are a new procedure for enumerating possible insertions and a new fitness criterion for choosing the best insertion. All new heuristics have quadratic worst-case computational complexity except for the critical-fit insertion heuristic which has a cubic worst-case computational complexity. The efficiency and effectiveness of the proposed heuristics is demonstrated by comparing their empirical performance on a standard benchmark data set against other published approaches.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science (General)
Authors
Krzysztof Fleszar,