The Swarthmore Computer Science department has extremely high demand for courses. As a result, not everyone who wants to take certain courses in the department is able to.
The current system to make these decisions is essentially manual, and is extremely difficult and time consuming. This application provides a (relatively) easy to use interface to allow people without experience in Linear Programming to decide who gets accepted into the course.
You can define an arbitrary number of constraints that can be either numeric (such as "I can only have N students in the class", or "I want at least M first year students to take the course"), or percentage (such as "I want men to make up at most 60 percent of the course").
The parser will automatically take this plaintext file and formulate it into a linear program, which is then solved using scipy's linear program solver.
This is very much a work in progress, but I hope to extend this to allow a richer set of goals. Currently, goals are prioritized one after another (meaning you must fully optimize the first goal before you can optimize the next). It might be interesting to allow more variability with this. In addition, I'd like to have a seperate .csv parser to automatically populate the attributes array.
Hopefully, the department will find this code to be of use :)