International Timetabling Competition 2011: An Adaptive Large Neighborhood Search algorithm

Adaptive Large Neighborhood Search was rst developed as a metaheuristic for the class of Vehicle Routing Problems (Pisinger and Ropke (2005); Ropke and Pisinger (2006)). It has been applied for few other problem classes as well, including Project Scheduling (Muller (2009, 2010)), Lot-sizing (Muller and Spoorendonk (2011)), Optimal Statistic Median Problem (Katterbauer et al (2012)). Recently we have developed a framework based on ALNS for solving combinatorial optimization problems (written in C# 4.0). This framework is part of the commercial product Lectio, where it is used to solve various practical timetabling problems, see Kristiansen et al (2011); Sorensen and Stidsen (2012) and Kristiansen and Stidsen (2012). The pseudo code for a general ALNS algorithm is given in Algorithm 1.