We initially generated 180 instances grouped into 18 classes according to the number of symbols (n) and the maximum length of the sequences (TMAX). We considered instances with 5, 7, 9, 11, 13, and 15 symbols, and TMAX equal to 2n, 3n, and 4n. For each pair (n, TMAX), ten distinct instances were generated by assigning random priorities to each symbol in the interval [1, 2n]. The name of the instances follows the format ins_<n>_<TMAX>.txt. The instances can be downloaded from the link instances.zip. The content of the files is organized as follows:
<n>
<TMAX>
<c1 c2 c3 ... cn>
where TMAX can assume the values 2, 3, or 4, meaning maximum lengths equal to 2n, 3n, and 4n, respectively.
To test the performance limits of the proposed methods, we created new classes by varying n ∈ {5, 7, ..., 15, 20, ..., 50} and TMAX ∈ {2n, 3n, ..., 10n, 15n, ..., 40n}, with 10 instances each. These instances can be downloaded from the link largeInstances.zip.