pyjobshop / fjsplib Goto Github PK

This issues discusses how to include setup times in the FJSPLIB format.

To include:

• #6
• #5

Keep in mind that we need to support:

• Open shop
• Hybrid flow shop

FJSPLIB
https://people.idsia.ch/~monaldo/fjsp.html#ProblemInstances

New instances (?)
https://github.com/ai-for-decision-making-tue/Job_Shop_Scheduling_Benchmark_Environments_and_Instances

https://github.com/SchedulingLab/fjsp-instances

Da Col, G., & Teppan, E. C. (2022). Industrial-size job shop scheduling with constraint programming. Operations Research Perspectives, 9, 100249. https://doi.org/10.1016/j.orp.2022.100249

• Job shop instances

• Related: https://github.com/wsgisler/job-shop-scheduling

• https://github.com/SchedulingLab/jsp-instances

Dauzère-Pérès, S., Ding, J., Shen, L., & Tamssaouet, K. (2023). The flexible job shop scheduling problem: A review. European Journal of Operational Research. https://doi.org/10.1016/j.ejor.2023.05.017

contains an overview of benchmark instances.

There are several sets of FJSP benchmark instances that are widely used in the literature:
34 BRdata (Brandimarte, 1993), HUdata (Hurink et al., 1994), BCdata (Barnes and Chambers,
35 1996), DPdata (Dauzere-P ́eres and Paulli, 1997), DMUdata (Demirkol et al., 1998), KCdata
36 (Kacem et al., 2002a), Fdata (Fattahi et al., 2007), DAdata (Birgin et al., 2015), as well as
37 Ydata (Birgin et al., 2015), leading to a total of 467 instances of different sizes and with
38 different flexibility levels.

Test Instances for the Flexible Job Shop Scheduling Problem with Work Centers
Dennis Behnke1,∗ and Martin Josef Geiger1

also contains a benchmark set.

Lunadi 2020: online color printing instances
https://github.com/Willtl/online-printing-shop/tree/master

• Should be easy to solve this type of model.

Naderi, B., & Roshanaei, V. (2022). Critical-Path-Search Logic-Based Benders Decomposition Approaches for Flexible Job Shop Scheduling. INFORMS Journal on Optimization, 4(1), 1–28. https://doi.org/10.1287/ijoo.2021.0056

• New hard FJSP instances. Cant find the instances directly from the paper, but I know that Naderi et al. (2023) has them,

Morais, R., Bulhões, T., & Subramanian, A. (2023). Exact and heuristic algorithms for minimizing the makespan on a single machine scheduling problem with sequence-dependent setup times and release dates. European Journal of Operational Research. https://doi.org/10.1016/j.ejor.2023.11.024

• 1800 instances for single machine with SDST and release dates: https://github.com/rsobralm/SMRDSDST/tree/main
• Instances should be attributed to Ovacikt, I. M., & Uzsoy, R. (1994). Rolling horizon algorithms for a single-machine dynamic scheduling problem with sequence-dependent setup times. International Journal of Production Research, 32(6), 1243–1263. https://doi.org/10.1080/00207549408956998
  • What's strange about these instances is that they include information about the number of machines but they're only used to create single machine instances.

This issue discusses how to include arbitrary precedence graph data in the FJSPLIB format.

FJSPLIB traditionally uses zero-based indexing. Do we want to adhere to this choice or switch to one-based indexing like https://github.com/SchedulingLab/fjsp-instances did?

I want to convert job shop scheduling instances in FJSPLIB format for benchmarking. That should be relatively easy to do.