Iterated local search for the team orienteering problem with time windows

  • Authors:
  • Pieter Vansteenwegen;Wouter Souffriau;Greet Vanden Berghe;Dirk Van Oudheusden

  • Affiliations:
  • Centre for Industrial Management, Katholieke Universiteit Leuven, Celestijnenlaan 300A-bus 2422, 3001 Leuven, Belgium;Centre for Industrial Management, Katholieke Universiteit Leuven, Celestijnenlaan 300A-bus 2422, 3001 Leuven, Belgium and Information Technology, Katholieke Hogeschool Sint-Lieven, Gebroeders Desm ...;Information Technology, Katholieke Hogeschool Sint-Lieven, Gebroeders Desmetstraat 1, 9000 Gent, Belgium;Centre for Industrial Management, Katholieke Universiteit Leuven, Celestijnenlaan 300A-bus 2422, 3001 Leuven, Belgium

  • Venue:
  • Computers and Operations Research
  • Year:
  • 2009

Quantified Score

Hi-index 0.01

Visualization

Abstract

A personalised electronic tourist guide assists tourists in planning and enjoying their trip. The planning problem that needs to be solved, in real-time, can be modelled as a team orienteering problem with time windows (TOPTW). In the TOPTW, a set of locations is given, each with a score, a service time and a time window. The goal is to maximise the sum of the collected scores by a fixed number of routes. The routes allow to visit locations at the right time and they are limited in length. The main contribution of this paper is a simple, fast and effective iterated local search meta-heuristic to solve the TOPTW. An insert step is combined with a shake step to escape from local optima. The specific shake step implementation and the fast evaluation of possible improvements, produces a heuristic that performs very well on a large and diverse set of instances. The average gap between the obtained results and the best-known solutions is only 1.8% and the average computation time is decreased with a factor of several hundreds. For 31 instances, new best solutions are computed.