Benchmark instances for the preemptive single machine scheduling of equal-length jobs minimizing Total Weighted Tardiness
Description
The benchmark is an attachment to the paper "The benefit of preemptions in single machine scheduling of equal-length jobs for Total Weighted Tardiness" article by Artem Fomin and Boris Goldengorin. Our benchmark provides over a million 1|pmtn;pj=p;rj|∑wjTj problem instances. The instances' size is up to np=700, where n is the number of jobs and p is the common processing time with n in [10, 350] and p in [2, 20]. For each instance we provide the optimal objective function value, corresponding optimal schedule and time in which it was found by the algorithm described in the paper. The instances in this benchmark are the same as in our previous dataset for the Total Weighted Completion Time (TWCT) problem with due dates added. Some of the bigger instances are omitted, as they could not be solved by our model in time. The previous dataset is available at: https://data.mendeley.com/datasets/nrkx7467tf For a more detailed dataset description, see the "readme_twt.txt" file. For an even more detailed description refer to the mentioned article. For your convenience, we also provide a small version of the benchmark with only the first 100 instances for each problem size. It will be much easier to handle than the full dataset with more than a million instances, each in a separate file.