MAPT Dataset
Mandarine Academy Professional Timetabling (MAPT) is a real-world dataset suggested to solve Professional Timetabling Problems (PTPs). A rather under-exploited category of the overall Timetabling Problems. However, we believe it can still be applied to traditional problems (education, health, etc.) as a helpful benchmark dataset to assist researchers in comparing different methods. Compared to conventional educational datasets such as (ITC2007), MAPT proposes richer features inspired by real-world data to provide insight into corporate training logistics and timetabling complexities.
Dataset Description
Two different kinds of records are in MAPT
- 1. Input files that are used as testing the approach. Having 3 other groups (Small, Medium, and Large) each in 20 different instances, totaling 60 test sets, to simulate different real-world scenarios.
- 2. Second records are training files that include the information of each entity involved in the scheduling process. Such as training, teachers, rooms, locations, etc. Note that both records have no redundant values.
Test files (Input/)
We provide test files used in our experiments. To simulate different real-world scenarios, we created varying complexities (number of events and available time window). Rather than using synthetic data, we use historical data for experiments. Problem instances were created using a total of 152 different training. Table.1 offers an overview of our test instances. Note here that there are no duplicate courses inside each test instance.
In total, 3 different problem sizes (small, medium, and large) are provided. For each problem size, the number of sessions, training, and planning window is increased. This was implemented to simulate how the algorithms will behave when having a larger number of events to satisfy. Note that the duration is expressed in days and each size has 20 different test instances.
- Folder mapt_sm: The full MAPT (Small) test instances (20).
- Folder mapt_md: The full MAPT (Medium) test instances (20).
- Folder mapt_md_2: The full MAPT (Medium) with limited time window test instances (20).
- Folder mapt_lg: The full MAPT (Large) test instances (20).
- Folder mapt_lg_2: The full MAPT (large) with limited time window test instances (20).
Training files (Dataset/)
- trainings.csv: This file has essential information about all training used in the planning process (ID, Type, Duration, etc.)
- training_skills.csv: This file provides a list of required skills for each training.
- training_rooms.csv: Associated rooms to training.
- training_devices.csv: Associated devices to training.
- training_animators.csv: Associated teachers to training.
- training_skilled_animators.csv: Associated skilled teachers to training
- sequences.csv: This file has essential information about all sequences used in the planning process (ID, Name, Duration, etc.).
- sequence_rooms.csv: Associated rooms to sequences.
- sequence_devices.csv: Associated devices to sequences.
- sequence_animators.csv: Associated teachers to sequences.
- sequence _skills.csv: This file provides a list of required skills for each sequence.
- resource_rooms.csv: This file has essential information about all rooms (ID, capacity, location, etc.).
- resource_devices.csv: This file has essential information about all equipment (ID, type, etc.).
- resource_animator.csv: This file has essential information about all teachers (ID, type, etc.).
- animator_skills.csv: This file provides a list of skills associated with each teacher.
- rooms_unavailability.csv = This file provides reservation information (dates, times, location, etc.) made for each room.
- device_unavailability.csv = This file provides reservation information (dates, times, location, etc.) made for each piece of equipment.
- animator_unavailability.csv: This file provides reservation information (dates, times, location, etc.) made for each teacher.
- localization.csv: these files have information (id, city, country, etc.) about where events can take place.
Initial Solutions (Initial Solutions/)
- Initial solutions were made available using our proposed constructive heuristic found in our work. These can be used directly as input to benchmark approaches.
- Folder mapt_sm: Initial solutions using the first input file sm_0.csv.
- Folder mapt_md: Initial solutions using the first input file md_0.csv.
- Folder mapt_md_2: Initial solutions using the first input file md_2_0.csv.
- Folder mapt_lg: Initial solutions using the first input file bg_0.csv.
- Folder mapt_lg_2: Initial solutions using the first input file bg_2_0.csv.
Final non-dominant solutions (Non-dominant solutions/)
We provide final objectives and directions. We included NSGAII and NSGAIII results using only 3 Objectives.
- Folder mapt_sm: Final objectives values using the first input file sm_0.csv.
- Folder mapt_md: Final objectives values using the first input file md_0.csv.
- Folder mapt_lg: Final objectives values using the first input file bg_0.csv.
Citing
When referring to the data, please cite the following paper:
@data{DVN/A4JU5E_2022,
author = {Hafsa, Mounir},
publisher = {Harvard Dataverse},
title = {{Mandarine Academy Professional Timetabling Dataset}},
UNF = {UNF:6:eE06QeDwLB06zCXDKuP1tQ==},
year = {2022},
version = {V1},
doi = {10.7910/DVN/A4JU5E},
url = {https://doi.org/10.7910/DVN/A4JU5E}
}
Request for Dataset
Use this form to request the dataset. Only non-commercial research is permitted on this dataset. This is a manual process so please allow at least two weeks to hear from us with the data sharing agreement.