Rectangle Packing Evaluation of the Region Packing Problem

Data Structure

The dataset consist of three classes with 200 packing problems each in '.elkt' format.
Additionally, I included generation files '.elkr' for these classes as well as '.grana' evaluation files to create the data using the KIELER SCCharts tool.

The results are in CSV files containing the results of running the algorithms.
The CSV files contain the following evaluation data:

...

- File: The name/number of the packing problem
- Area (Width): The width
- Area (Height): The height
- Area (Product): The area
- Aspect Ratio
- Whitespace (Top-Level)
- Whitespace (Hierarchy): Same as top-level
- Scale Measure (Linear): Desired aspect ratio
- Scale Measure (Scale Measure): The scale measure
I evaluated the following strategies using these datasets:
- diss_eval_R: Rectpacking heuristic
- diss_eval_R2: Rectpacking heuristic with 2 iterations for compaction
- diss_eval_RB: Rectpacking heuristic with trying box algorithm first
- diss_eval_RR: Rectpacking heuristic with row height reevalation
- diss_eval_RX: All optimization together with as many compaction iterations as an improvement can be made (10 is enough).
- rectpacking_cplex: Results of the optimization problem capped at 1 hour execution time

Additionally, sccharts-packings.zip holds the same evaluation for rectangle packing problems posed by SCCharts models (which cannot be included in this dataset).
Here, the following strategies were evaluated:
- box-new: The simple heuristic, the box algorithm
- cplex-new: Again uses cplex to solve the packing as a optimization problem
- rectpacking: Using the rectpacking heuristic
- rectpacking-2: Rectpacking heuristic with 2 iterations for compaction
- rectpacking-box: Rectpacking heuristic with trying box algorithm first
- rectpacking-rowHeight: Rectpacking heuristic with row height reevalation
- rectpacking-X: All optimization together with as many compaction iterations as an improvement can be made (10 is enough).
- ".grana" file to evaluated the data using the KIELER SCCharts tool