Solving the course timetabling problem with a hybrid heuristic algorithm

Abstract

The problem of curriculum-based course timetabling is studied in this work. In addition to formally defining the problem, we present a hybrid solution algorithm (Adaptive Tabu Search-ATS), which is aimed at minimizing violations of soft constraints. Within ATS, a new neighborhood and a mechanism for dynamically integrating Tabu Search with perturbation (from Iterated Local Search) are proposed to ensure a continuous tradeoff between intensification and diversification. The performance of the proposed hybrid heuristic algorithm is assessed on two sets of 11 public instances from the literature. Computational results show that it significantly improves the previous best known results on two problem formulations. © Springer-Verlag Berlin Heidelberg 2008.