Project title Dynamic Course Allocation
Duration  10 weeks

In many countries around the world, the allocation of courses within universities is realized through a centralized process. Examples include Brazil, Costa Rica, and Singapore, among others.  Typically, students are prioritized by a score, and express preferences over courses in a particular period (semester). Then, the Registrar Office (RO) uses an algorithm to assign students to courses, such that preferences, priorities and the number of seats in each course are respected. An additional feature of some assignment process is that students’ priorities are endogenous and evolve over time. For example, students are prioritized by their score in their admission test in the first period (t=1), but their priority in period t>1 depends on their GPA in the last period. When the RO uses this system, students have to be strategic in the way they complete their plan of studies, which can delay their graduation time, or in the worst scenario, can force them to change their major or walk away from the university.

The current project aims to analyse the properties of the endogenous priority assignment (EPA) with respect to welfare, fairness and stability. From the understanding of the EPA’s failures, we will design a new mechanism and will provide an algorithm to implement it.

Expected outcomes & deliverables

Applicants will be introduced in the market design literature. This literature deals with two-sided problems in which prices cannot mediate to assign scarce resources to alternative candidates. Examples include adoption and foster systems, kidney exchange, school choice, allocation of doctors to hospitals and teachers to schools, and allocation of schedules to nurses, among many others. Some of these topics are of high interest for policy makers in Australia, and offer the possibility to develop a wide research agenda.

Applicants will also trained on how to approach a very applied problem with the rigour of microeconomic techniques. Specially those related to game theory, information economics and discrete optimization.

Applicants will also learn how to develop algorithms for matching problems in Python, Matlab or any related language.

We expect to the applicants to be able to do a review of literature, understand the essence of the problem, work with easy examples, write algorithms for those examples, and read proofs when necessary.

Student qualities

The project is suitable for a student with a background in economics, applied mathematics or computer science. Preferably (but not necessary), students who have taken intermediate (advanced) microeconomics as undergraduates. Students with abilities to program in Python or Matlab or a similar language, will be given priority.

Primary supervisor

Dr. Allan Hernandez-Chanto

Further information Candidates interests in applying to the project can write to if they have specific inquiries about the project.