The Dorm Assignment Optimizer (DAO) project focuses on using simulated annealing to optimize dorm room assignments. Juxtaposing the choices and chances underlying the process of assigning dorm rooms, in this thesis, we present the essence of the problem, our solution, and results of this two year project.
Assigning dorm rooms can be viewed as a collection of choices made by the students and the Housing Reservations staff, and a series of chances taken by our algorithm to produce a guaranteed local optimal assignment of students to beds. Choices and chances is a useful distinction even though there will be some chance in the problem and choice in the solution.
DAO production results have demonstrated superior performance in terms of student placement and use of resources during the last three semesters of use at UNM, over the previously used vendor-supplied mechanism.