RACE
RACE is a basic research project which studies the automated selection and parameterization of heuristic optimization methods.
The Problem:
Routing a fleet of vehicles optimally is a complex and difficult task. Heuristic methods, especially Metaheuristics, often provide good solutions that can be implemented as part of a computerized planning system, but applying them to a given setting in the best possible way is a very difficult task in itself which currently requires a lot of effort and expertise.
Our Goal:
We want to explore the impact of the choice of methods and parameters, and examine the possibility of automating this selection.
Our Approach:
In the RACE project, we will:
- establish a common framework in which various vehicle routing problems and heuristic solution methods can be programmed;
- find mathematical methods to evaluate and compare heuristic solutions;
- combine these elements to automate key areas of solution development.
Benefit:
The know-how acquired in the RACE project will help us develop better solutions for vehicle routing problems.
We will be able to offer tailored practical solutions even for industries with very specific requirements, such as ambulance fleets or waste disposal, and do so with less effort than previously required.
