EvoWebReg: Web-Based Course Registration and Optimization of Student Personal Schedules with Evolutionary Algorithms

Problem Description – Case Study

According to Carter and Laporte (1998) the course scheduling problem can be decomposed into the following sub-problems: course timetabling (assign courses or course sections to time periods satisfying some requirements and/or constraints), teacher assignment (assign teachers to courses maximizing a preference function), classroom assignment (events must be assigned to specific rooms to satisfy some criteria), and student scheduling (arises when courses are taught in multiple sections). The course timetabling systems are also distinguished as Master Timetables or Demand Driven, with their primary difference lying in the sequence in which the various sub-problems are being solved.

In this paper we consider the student scheduling problem with the Master Timetable approach, adapted to the requirements and characteristics of the Dept. of Informatics of ATEITh. Although we discuss a specific case study, most methods are general enough and should be applicable to modeling and solving other instances of the problem.

In our case study, a semester is the academic time unit. Course timetables and student sectioning must be done in the beginning of every semester. Studies last for 8 semesters. In the first seven semesters, the students have to pass a number of courses. In the eighth semester students have to elaborate a thesis and also complete a six month practice at a company (industrial placement).