Dataset used in article "A genetic algorithm approach for the two-dimensional variable-sized stock problems"

The dataset presented is part of the one used in the article “A genetic algorithm approach for the two-dimensional variable-sized stock problems” by Paula Terán-Viadero, Antonio Alonso-Ayuso, Maria Antónia Carravilla, F. Javier Martín-Campo and José Fernando Oliveira, submitted for publication (2025). In the paper mentioned above, a genetic modelling framework for a variant of the Two-Dimensional Variable-Sized Cutting Stock Problem (2DVSCSP) is presented. The objective of the framework is to simultaneously determine the stock sizes to be produced and the cutting patterns assigned to them. Each stock size is defined by its width and length, while each cutting pattern is characterised by the item types included and the number of rows of each item type. The computational experiments analyse different production settings by limiting: The maximum number of different widths allowed in the solution. The maximum number of different stock sizes allowed. Two stock-dimension settings are considered: Discrete widths with continuous lengths. Continuous widths with continuous lengths. Additionally, patterns containing up to two or three item types are analysed. The dataset presented here includes the input instances and the computational results obtained for the different experimental configurations analysed in the paper. The repository contains: The input data files defining the item dimensions, demands, stock length bounds, and available widths. The output files containing the stock sizes and cutting patterns generated by the framework. Execution statistics associated with the BRKGA-IPR metaheuristic used in the solution process. For each experimental configuration, five different random seeds were used. The dataset is organised according to: The stock-dimension setting. The maximum number of item types per pattern. The maximum number of different widths allowed. The maximum number of different stock sizes allowed.