Generating feasible university student schedules via graph theory concepts -

Abstract

This thesis aims to provide an effective method that helps university students in registering several classes having multiple sections each. With many classes to register (reaching 6 or 7 classes) and several sections per class (reaching 8 or 9 sections) this becomes a scheduling problem that is tedious to solve. The difficulty stems from finding available options (schedules) that comply to a mixture of hard constraints imposed by the university and soft constraints related to student’s preferences. This thesis describes the design and implementation of a constraint-based personal student course scheduler that utilizes graph theory. A special graph is constructed which consists of special vertices (referred to as super nodes) that represent the courses offered by the university. Each super node has sub-nodes representing sections of the class. The connectivity is defined among super nodes. Two super nodes are connected by an arc if they have no time conflict. Sub-nodes (sections) that belong to connected super-nodes can be part of the student schedule. This graph theory methodology allows generating all possible schedules for a student accounting for various constraints. We apply our method to generate student course schedules based on the course offerings of the American University of Beirut (AUB). This is developed using a user-friendly PHP application which runs online, and can be easily connected to a university student information system.