A simulated annealing algorithm for university course timetabling considering travelling distances

Abstract

This paper addresses the curriculum-based course timetabling problem CB-CCT with extra travelling distance constraints using an improved simulated annealing algorithm. The motivation of considering the new constraints is that the travelling distances between two consecutive lectures attended by the same group of students must be controlled into a certain range in order to facilitate on-campus transportation. We propose a two-stage solution method, which circumvents the hard constraints violations firstly, and minimises the soft constraints violations subsequently. To obtain satisfactory solutions, an improved simulated annealing algorithm with memory function ISAM is adopted, which incorporates a variety of space adjustment and time adjustment strategies as the local search module. We study the impact of the initial solution, the annealing strategy, and the search depth of ISAM on the final solution quality. The computational results show that the improved simulated annealing algorithm can solve the course timetabling problem effectively.