Scheduling solution for my needs to match persons optimally to assignments based on 4 levels of preferences.
To get started, type python scheduling_solution.py --help.
An example for the input CSV file is available on the repository tmp.csv.
A solution for scheduling hell!
Create a CSV with the following properties:
- cell A1 = 'Assignment'
- cell A2 = 'Counts'
- cells A3-A[#] being the names of the entities that require assignments. For instance, "Section 1", "Section 2", etc.
- cell B1 = 'Counts'
- leave cell B2 blank
- cells B3-B[#] being the number of assignments to be made to the corresponding entity in A3-A[#]. For instance, Section 2 requires 3 persons assigned to it.
- cell C1 should contain the name of a person (who needs to be assigned an entity)
- cell C2 should contain the count of entities to be assigned to this person
- cells C3-C[#] should contain the preferences of the person to be assigned to the corresponding entity. The prefernce must be {1, 2, 3, 4}, where 1 is "most preferred" and 4 is "cannot make it at all."
- follow the same thing for more persons as column C.
This script can then be run by providing the input csv name and an output csv name.
More options:
-
It is possible to skip column B and row 2 (the counts), and let the program assume that all counts are 1.
-
It is possible to specify the minimum/maximum counts and let the program generate assignments that are allowed to be over/under the counts provided in the CSV. Common use case: There are 4 homework party slots, and 18 TAs. Each slot needs a minimum of 4 people, but the remaining 2 can be assigned freely to the best preferred slot. In this case, this script can be run with --bound lower, specifying that the counts in the CSV are a lower bound.