„Vehicle routing“ describes a class of notoriously difficult problems arising in the context of the physical distribution of goods. Besides the minimizing of cost, other aspects have to be taken into consideration such as the maximization of the provided service with respect to agreed delivery dates. This results in multi-objective formulations of the problem, in which a compromise plan has to be found balancing the different aspects of the problem.
In our research, we develop approaches for vehicle routing problems under multiple objectives. This implies the proposition of solution techniques for the identification of efficient solutions, as well as the development of decision-aiding methods, helping the human planner to articulate his/ her preferences.
Our work in multi-objective vehicle routing goes back to the early development of a decision support system based on genetic algorithms. In this approach, the Pareto-front is first approximated by the implemented heuristic, and then shown to the decision maker.
In brief, the planner is therefore able to select (i) a solution minimizing the costs, (ii) a solution minimizing the time window violations, or (iii) any other efficient solution obtained by the heuristic search algorithm.
The developed software participated in the finals of the European Academic Software Award 1998, held at The University of Oxford, Keble College, Oxford, UK.
Letzte Änderung: 21. November 2017