Hybrid artificial intelligence technique for solving large, highly constrained timetabling problems

Abstract

Timetabling problems are often hard and time-consuming to solve. Profits from full automatization of this process can be invaluable. Although over the years many solutions have been proposed, most of the methods concern only one problem instance or class. This paper describes a possibly universal method for solving large, highly constrained timetabling problems from different areas. The solution is based on evolutionary algorithm's framework, with specialized genetic operators and penalty-based evaluation function, and uses hyper-heuristics to establish its operating parameters. The method has been used to solve three different timetabling problems, which are described in detail, along with some results of preliminary experiments.