PhD candidates at the Econometric Institute of Erasmus School of Economics Bart van Rossum and Rick Willemsen have been successful in an optimisation challenge dedicated to a 3D truck loading problem provided by Renault Group. Every two years, the French Operational Research and Decision Support Society (ROADEF) organises this challenge together with the European Operational Research Society (EURO).
Best performing team in the junior category
The challenge started in July 2022 and consisted of three rounds of competition. In the final phase in July 2023, young researchers Bart van Rossum and Rick Willemsen came out on top as the best performing team in the junior category. Overall, they finished in 5th place among 51 registered teams. The first place in the junior category, dedicated to researchers who have not yet completed their PhD, is rewarded with a cash prize of 5,000 euros.
An industrial-sized truck loading problem
The 2022 ROADEF challenge offered participants the opportunity to work on a large-scale truck loading problem constructed in collaboration with industrial partner Renault Group.
6000 trucks deliver parts from suppliers to plants
Renault’s supply chain spans over 40 plants in 17 countries and 1,500 suppliers. Every week, 6,000 trucks deliver parts from suppliers to plants. The filling rate of these trucks is critical since the annual inbound transportation budget exceeds several hundred million euros.
Pack a set of items from suppliers into stacks
The objective is to pack a set of items from suppliers into stacks and to pack the stacks into trucks that deliver to the plants, to minimise the number of trucks used and the inventory in the plants due to early deliveries. Renault provided hundreds of large-scale instances for the challenge, containing up to 260,000 items and 5,000 planned trucks.
Winning algorithm built on techniques from own research
To compete in the challenge, Van Rossum and Willemsen developed an optimisation algorithm for the above truck loading problem. While this industrial application features many complicated requirements and practical difficulties, Bart and Rick were able to use techniques from their own research in constructing an efficient solution method. The winning algorithm relies on column generation, an operations research technique commonly applied to large-scale problems. In order to deal with instances of ever-increasing size, many heuristic speed-ups were added throughout the three rounds of competition.
- More information
For more information, please contact Ronald de Groot, Media & Public Relations Officer at Erasmus School of Economics: rdegroot@ese.eur.nl, +31 6 53 641 846.