Office Operations Scheduling

Scheduling script for CSSU lounge (BA2250) office operations, by solving a Constraint Satisfcation Problem (CSP). Yes, you should take CSC384.

Usage

1. Prepare <input.csv> (with headers) where each cell is the list of people available to serve that time. Here is an example:

   Time/Day,Tue,Wed,Thu
   9:00 AM,A|B,,A|B
   10:00 AM,,,
   11:00 AM,,C,C
   12:00 PM,,,
   1:00 PM,A|B,A|B|C,
   

2. Check/edit hyperparameters.

   # sort by least available first
   people_sorting = lambda p: (len(availability[p]), float("inf")
                              if p not in required else -required[p])
   
   # centres-of-gravity
   centre = ((len(content[0]) - header[0]), (len(content) - header[1]) // 2)
   
   # centre around day=1, time=2
   time_prefs["A"] = lambda t: (abs(1 - t[0]), abs(2 - t[1]))
   
   # day doesn't matter/time doesn't matter
   time_prefs["B"] = lambda t: (float("inf"), abs(centre[0] - t[0]))
   time_prefs["C"] = lambda t: (abs(centre[1] - t[1]), float("inf"))
   
   # require two slots per person, including custom slots for "A"
   required = dict([(p, 2) for p in availability])
   required["A"] = 3

3. Run the script to produce <output>.md.

   python scheduler.py <input.csv> <output.md>

   The resulting file looks like this.

   | Time/Day | Tue | Wed | Thu |
   | -------- | --- | --- | --- |
   | 9:00 AM  | A   |     | A   |
   | 10:00 AM | A   |     |     |
   | 11:00 AM |     | C   | C   |
   | 12:00 PM |     |     |     |
   | 1:00 PM  | B   | B   |     |