An instance of the skiving stock problem is given by a threshold length and a set of item types specified by their length and frequency for . A set of instances consists of several instances with the same threshold and number of item types.

Each file in the data directory contains a particular set of instances. The format is as follows:

• the first line contains a number which is equal to the number of instances in the set,
• each following line represents an instance written in the form , where all values are integers separated by a tab character.

The directories A1 to A3 and B contain artifical instances generated according to the method described in [1]. The directories C1 to C3 contain other benchmark datasets from the literature. Details can also be found in Section 4 of [1].

[1] Martinovic, J., Delorme, M., Iori, M., Scheithauer, G., & Strasdat, N. (2020). Improved flow-based formulations for the skiving stock problem. Computers & Operations Research, 113, 104770.