The Naval Postgraduate School currently uses a time consuming manual process to generate course schedules for students and professors. Each quarter, the process of timetabling approximately 2000 students into nearly 500 courses takes up to 8 weeks. This thesis introduces an automated timetabling algorithm using Probability Collectives (PC) theory. PC Theory is an agent based approach that utilizes Collective Intelligence (COIN) to solve optimization problems by using a collection of agents attempting to achieve a single goal. The algorithm was tested on a set of data provided by the organizers of the 2007 International Timetabling Competition. The algorithm provided valid timetables for every problem instance and successfully scheduled between 70% and 91.6% of all student course requests.
APPROVED FOR PUBLIC RELEASE
Keywords relating to this report:
Email This Abstract
