Data for: "Generating textual explanations for scheduling systems leveraging the reasoning capabilities of Large Language Models"

This data set was created for the paper titled 'Generating textual explanations for scheduling systems leveraging the reasoning capabilities of Large Language Models'. The paper looks into the capabilities of Large Language Models (GPT4 and LLama 3.1) in analysing and interpreting benchmark job shop schedules in tabular form, containing different query categories, namely:

1. Swap - Queries that entail exchanging jobs on the same machine as well as different machines.
2. Increase - Queries around modifying the duration times of jobs on different machines.
3. Decrease - Queries where job duration times are decreased on different machines.

These questions were applied to different variations and sizes of benchmark job-shop schedules. The variations are:

1. Seven different schedule variations containing 15 jobs on 15 machines (15j x 15m).
2. One variation of 20 jobs on 15 machines (20j x 15m).
3. One variation of 20 jobs on 20 machines (20j x 20m).
4. One variation of 30 jobs on 20 machines (30j x 20m).

There are two main folders within this data set that contain the questions, answers and scores for each schedule.

These are:

1. GPT - contains 4 folders
a. GPT 15_15 - Contains seven folders, one for each variation of 15j x 15m schedule, with each containing three spreadsheet files for each query category
b. GPT 20_15 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 20j x 15m.
c. GPT 20_20 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 20j x 20m.
d. GPT 30_20 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 30j x 20m.

2 Llama - contains 4 folders
a. Llama 15_15 - Contains seven folders, one for each variation of 15j x 15m schedule, with each containing three spreadsheet files for each query category
b. Llama 20_15 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 20j x 15m.
c. Llama 20_20 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 20j x 20m.
d. Llama 30_20 - Contains three spreadsheet files, one for each query category, containing questions for the schedule variation of 30j x 20m.